{"slug":"gaasher-karpathy","name":"karpathy","description":"Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor.","long_description":"---\nname: karpathy\ndescription: >\n Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the\n training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent\n proposes one change at a time, runs training in the user's env, keeps it only if the metric improves\n (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful\n adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing\n (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor.\ncompatibility: Requires Python 3.9+\nmetadata:\n version: \"0.1.0\"\n---\n\n# Karpathy Autoresearch\n\nThis is an experiment to have the LLM do its own research. You are a completely autonomous researcher:\nyou hack the training code with an idea, run it, keep the change if the metric improves and revert it if\nit doesn't, advancing a branch as you go — and you repeat **forever, until the human interrupts you.**\nThe artifact is the ``; the feedback signal is one scalar `` (lower is better,\ne.g. `val_bpb`) read from the run. Training runs in the user's own environment via `` — this\nskill installs nothing and imports nothing; it edits code, shells out, and reads the metric from the log.\n\n## When to use\nUse this to leave an agent running on a single training script, optimizing one scalar metric hands-off,\nwhere any improvement is kept and the loop never stops on its own. Default to broad freedom inside\n``; the only hard limit is that the run finishes within the budget without crashing. Not\nfor the analysis-first variant that reasons about the data before each edit (that is `ml-autoresearch`).\n\n## Setup\nResolve bindings interactively (load `loop.run.yaml` and skip if it already exists; else, on Claude Code\ninfer + recommend each via `AskUserQuestion`, otherwise ask as quoted prompts; write `loop.run.yaml`).\nThen **work with the user** to set up a fresh run:\n\n1. **Choose ``** — **branches** (one git commit per run; the original) or **snapshots**\n (one folder per run under `/`). Snapshots are safer on a dirty or gitignored tree;\n branches mirror Karpathy. Either is fully supported throughout the loop.\n2. **Open the run** — *branches:* agree on a run tag from today's date (e.g. `mar5`) and create the\n branch `git checkout -b autoresearch/` (it must not already exist; this is a fresh run).\n *snapshots:* no branch — each iteration gets its own `/iter/`.\n3. **Read the in-scope files** — the repo is small; read them for full context: the README, the\n **read-only** harness that defines the metric (the `` ground truth — do not modify), and the\n `` you will hack (model/optimizer/training loop).\n4. **Verify the env/data exists** — confirm `` can run (data shards, tokenizer, deps present).\n If not, tell the human the one command to prepare it (e.g. `uv run prepare.py`).\n5. **Initialize `results.tsv`** — create it with just the header row; the baseline is recorded after the\n first run. Leave it untracked (never commit it).\n6. **Confirm and go** — confirm the setup looks right, then kick off the experimentation.\n\n| binding | meaning | default | how to infer |\n|---|---|---|---|\n| `` | scalar to **minimize**; must be printed by the run (e.g. `val_bpb`) | — | grep the in-scope files / README for `val_bpb`, `val_loss`, `error`… |\n| `` / `` | the command that launches one training run | — | e.g. `uv run train.py`; from `pyproject.toml`/`.venv`/README |\n| `` | the file(s) you may hack — everything else is read-only | — | the training script(s); exclude data, the eval harness, configs you must not touch |\n| `` | where `results.tsv` (+ snapshots) live | `./sandbox` | — |\n| `` | `branches` (a git commit per run) or `snapshots` (a folder per run) | `snapshots` | `snapshots` is safer off a dirty/gitignored tree; `branches` mirrors Karpathy |\n| `` / `` | `time` (wall-clock) or `epochs`, and its value | — | the run's existing time/epoch setting |\n\n## The experiment loop\nEach experiment is one training run on a **fixed budget** (``/`` — wall-clock time or a\nfixed epoch count, excluding startup/compile). You launch it simply: `` (e.g. `uv run\ntrain.py`). Because the budget is fixed you don't need to worry about training time — every run gets the\nsame budget.\n\n**What you CAN do:** Modify `` — this is the only file you edit. Everything is fair\ngame: model architecture, optimizer, hyperparameters, training loop, batch size, model size, etc.\n\n**What you CANNOT do:** modify the read-only harness or the evaluation (the `` is the ground\ntruth); install new packages or add dependencies (use only what's already available).\n\nThe goal is simple: get the lowest ``. Since the budget is fixed, you don't need to worry about\ntraining time — it's always the budget. Everything is fair game: change the architecture, the optimizer,\nthe hyperparameters, the batch size, the model size. The only constraint is that the code runs without\ncrashing and finishes within the budget.\n\nVRAM is a soft constraint. Some increase is acceptable for meaningful `` gains, but it should not\nblow up dramatically.\n\n**Simplicity criterion:** All else being equal, simpler is better. A small improvement that adds ugly\ncomplexity is not worth it. Conversely, removing something and getting equal or better results is a great\noutcome — that's a simplification win. When evaluating whether to keep a change, weigh the complexity\ncost against the improvement magnitude. A 0.001 `` improvement that adds 20 lines of hacky code?\nProbably not worth it. A 0.001 `` improvement from deleting code? Definitely keep. An improvement\nof ~0 but much simpler code? Keep.\n\n**The first run:** Your very first run should always be to establish the baseline, so you will run the\ntraining script as is.\n\nCopy this checklist; **LOOP FOREVER:**\n- [ ] **1.** Look at the git state — the branch/commit you're on (snapshots: the next `iter/`).\n- [ ] **2.** Tune `` with one experimental idea by directly hacking the code.\n- [ ] **3.** Commit it (`git commit -am \"\"`; snapshots: copy `` into `iter/code_snapshot/` first).\n- [ ] **4.** Run the experiment: ` > run.log 2>&1` (redirect everything — do NOT use `tee` or let output flood your context).\n- [ ] **5.** Read the result: `grep \"^:\" run.log` (also grab peak memory if printed).\n- [ ] **6.** If the grep is empty the run crashed — `tail -n 50 run.log`, read the trace, fix if it's something dumb (typo/missing import), else give up after a couple of tries.\n- [ ] **7.** Record the result in `results.tsv` (do NOT commit it — leave it untracked).\n- [ ] **8.** If `` improved (lower), **advance** — keep the commit. \n- [ ] **9.** If it's equal or worse, `git reset` back to where you started (snapshots: restore from `code_snapshot/`). Go to 1.\n\nThe idea is that you are a completely autonomous researcher trying things out. If they work, keep. If\nthey don't, discard. And you're advancing the branch so that you can iterate. If you feel like you're\ngetting stuck in some way, you can rewind but you should probably do this very very sparingly (if ever).\n\n**Output format.** When the run finishes it prints a summary; the exact lines depend on what the user's\nscript prints (`` is whatever you bound — `val_bpb`, `val_loss`, a perplexity, an error rate, …), e.g.:\n```\n: 0.997900\npeak_vram_mb: 45060.2\nnum_params_M: 50.3\n```\nThe numbers vary by machine since each run stops at the budget. Extract the metric with\n`grep \"^:\" run.log`.\n\n**Timeout.** A run should take ~its budget plus a little eval overhead. If a time-gated run exceeds\n`2× ` minutes, kill it and treat it as a failure (discard and revert).\n\n**Crashes.** Use judgement: something dumb and easy (a typo, a missing import) — fix it and re-run; an\nidea that's fundamentally broken — skip it, log `crash` as the status, and move on.\n\n## results.tsv (logging results)\n`/results.tsv`, tab-separated (NOT comma-separated — commas break in descriptions). Header\n+ 5 columns: the git commit (short, 7 chars; or `iter` in snapshots mode), `` (e.g. `1.234567`,\nor `0.000000` for a crash), peak memory in GB (`.1f`, `peak_vram_mb`/1024; `0.0` for a crash), `status`\n∈ {`keep`, `discard`, `crash`}, and a text description of what the experiment tried.\n```\ncommit\t\tmemory_gb\tstatus\tdescription\na1b2c3d\t0.997900\t44.0\tkeep\tbaseline\nb2c3d4e\t0.993200\t44.2\tkeep\tincrease LR to 0.04\nc3d4e5f\t1.005000\t44.0\tdiscard\tswitch to GeLU activation\nd4e5f6g\t0.000000\t0.0\tcrash\tdouble model width (OOM)\n```\nReport the **best** run when interrupted, not necessarily the last.\n\n## Constraints\n- **Only edit ``.** The read-only harness that produces `` is the ground truth —\n editing it (or the eval) would corrupt the signal the loop is scored against.\n- **One change per iteration**, so each `` delta is attributable to a single idea.\n- **Run in the user's env via ``.** Install nothing, add no dependencies — shell out and read\n the log. Always redirect output to `run.log`; never `tee`, never flood your context.\n- **Don't commit `results.tsv`** — leave it untracked. `/` is self-contained (no `../`).\n\n## Stops — NEVER STOP\nOnce the experiment loop has begun (after the initial setup), do NOT pause to ask the human if you should\ncontinue. Do NOT ask \"should I keep going?\" or \"is this a good stopping point?\". The human might be\nasleep, or gone from a computer and expects you to continue working indefinitely until you are manually\nstopped. You are autonomous. If you run out of ideas, think harder — read papers referenced in the code,\nre-read the in-scope files for new angles, try combining previous near-misses, try more radical\narchitectural changes. The loop runs until the human interrupts you, period.\n","tagline":"Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if","category":"research","tags":["agent-skill"],"author":"gaasher","verified":false,"attribution":{"status":"registry_indexed","statusLabel":"Registry indexed","shortLabel":"REGISTRY INDEXED","sourceLabel":"github fast track","sourceDetail":"gaasher/Agent-Loop-Skills","creatorName":"gaasher","creatorUrl":"https://github.com/gaasher","sourceUrl":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","indexedBy":"OpenAgentSkill community index","claimUrl":"https://www.openagentskill.com/skills/gaasher-karpathy#claim-this-skill","claimCta":"Claim this skill","trustNote":"This listing was indexed from public sources and is not marked official until a maintainer claim is approved.","publicNote":"Attribution links to the public repository or creator profile. Creators can claim the listing to update ownership signals."},"stats":{"stars":163,"forks":19,"verified_installs":0,"successful_runs":0,"total_outcomes":0,"rating":0,"review_count":0,"quality_score":35.6},"quality":{"score":63,"tier":"promising","label":"Promising","summary":"Useful candidate, but compare it with alternatives before adopting.","signals":[{"label":"GitHub stars","value":"163","tone":"neutral"},{"label":"Freshness","value":"2mo ago","tone":"positive"},{"label":"Install ready","value":"Yes","tone":"positive"},{"label":"License","value":"MIT","tone":"neutral"}],"warnings":[]},"trust":{"version":"trust-score-v5","score":65,"base_score":73,"outcome_confidence":0,"tier":"review","label":"Sandbox only","summary":"Useful candidate with missing or mixed trust signals. Keep it in an isolated workspace until the outcome loop proves task fit.","recommendedAction":"Run only in a sandbox and compare close alternatives before using it for real work.","decision":{"install_policy":"human_review_before_install","auto_install_allowed":false,"human_review_required":true,"sandbox_first":true,"agent_action":"Compare alternatives before installing.","reasoning":["65/100 Trust Score v5","73/100 Trust Score v4 baseline","Needs more real agent outcomes before unattended install","Install path is available","Review before production"],"review_required_when":["The workspace contains production secrets, payments, private customer data, or irreversible actions.","The install command requests shell, network, credential, database, or broad filesystem access.","Outcome evidence is missing, recently failed, or required human review.","Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"]},"dimensions":[{"id":"github_adoption","label":"GitHub adoption","score":62,"weight":0.13,"status":"info","detail":"163 GitHub stars"},{"id":"repo_activity","label":"Stars/forks activity","score":57,"weight":0.08,"status":"warn","detail":"163 stars, 19 forks; issue activity unavailable in current metadata"},{"id":"maintenance","label":"Recent maintenance","score":88,"weight":0.14,"status":"pass","detail":"2mo since push"},{"id":"license","label":"License clarity","score":86,"weight":0.09,"status":"pass","detail":"MIT"},{"id":"documentation","label":"README/SKILL.md completeness","score":86,"weight":0.14,"status":"pass","detail":"Metadata includes enough usage and workflow context"},{"id":"dependency_risk","label":"Dependency/runtime risk","score":54,"weight":0.12,"status":"warn","detail":"command execution surface, credential or environment access"},{"id":"installability","label":"Install availability","score":92,"weight":0.1,"status":"pass","detail":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy"},{"id":"install_safety","label":"Install command safety","score":92,"weight":0.1,"status":"pass","detail":"standard package or runtime install path"},{"id":"permission_surface","label":"Permission surface","score":36,"weight":0.07,"status":"fail","detail":"secrets or environment access, shell or command execution"},{"id":"repository","label":"Repository evidence","score":86,"weight":0.04,"status":"pass","detail":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy"},{"id":"review_status","label":"Review status","score":88,"weight":0.05,"status":"pass","detail":"AI review data available"},{"id":"agent_outcomes","label":"Agent Proven outcomes","score":54,"weight":0.13,"status":"info","detail":"No agent outcome data yet"}],"checks":[{"status":"info","label":"GitHub adoption","detail":"163 GitHub stars"},{"status":"warn","label":"Stars/forks activity","detail":"163 stars, 19 forks; issue activity unavailable in current metadata"},{"status":"pass","label":"Recent maintenance","detail":"2mo since push"},{"status":"pass","label":"License clarity","detail":"MIT"},{"status":"pass","label":"README/SKILL.md completeness","detail":"Metadata includes enough usage and workflow context"},{"status":"warn","label":"Dependency/runtime risk","detail":"command execution surface, credential or environment access"},{"status":"pass","label":"Install availability","detail":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy"},{"status":"pass","label":"Install command safety","detail":"standard package or runtime install path"},{"status":"fail","label":"Permission surface","detail":"secrets or environment access, shell or command execution"},{"status":"pass","label":"Repository evidence","detail":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy"},{"status":"pass","label":"Review status","detail":"AI review data available"},{"status":"info","label":"Agent Proven outcomes","detail":"No agent outcome data yet"},{"status":"warn","label":"Ownership","detail":"No approved owner claim yet"},{"status":"pass","label":"OpenAgentSkill usage","detail":"1 views, 0 install copies"},{"status":"info","label":"Agent outcomes","detail":"No agent outcome data yet"}],"strengths":["AI review approved","Install path is available","Repository evidence is available","Recently maintained repository","Install command has no obvious high-risk pattern","Outcome loop is ready but needs first real agent run"],"warnings":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution","No real agent outcome reports yet","Human review required before unattended installation"],"evidence":{"stars":"163 GitHub stars","repoActivity":"163 stars, 19 forks","lastPushed":"2mo since push","license":"MIT","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","install":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","installSafety":"standard package or runtime install path","permissionSurface":"secrets or environment access, shell or command execution","documentation":"Strong README/SKILL.md context","agentOutcomes":"No agent outcome data yet","agentProvenScore":0,"outcomeConfidence":"0%","installPolicy":"human_review_before_install"},"installReadiness":{"ready":true,"command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","policy":"human_review_before_install","label":"Human review before install","notes":["Install path is available","Repository evidence is available","License is declared","No Agent Proven outcome evidence yet","2mo since push","Trust Score v5 requires review or sandbox-only use before install."]},"agentCompatibility":["Codex","Claude Code","Cursor","OpenAgentSkill CLI"],"riskSummary":{"level":"medium","label":"Review before production","notes":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"outcomeEvidence":{"total":0,"successes":0,"failures":0,"notRelevant":0,"successRate":null,"installAttempts":0,"riskBlocked":0,"setupRequired":0,"installSuccessRate":null,"avgOutputQuality":null,"avgTimeToUsefulMs":null,"productionOutcomes":0,"humanReviewRequired":0,"recentSuccessRate":null,"recentFailureRate":null,"uniqueAgents":0,"agentProvenScore":0,"agentProvenLabel":"Needs first agent run","lastOutcomeAt":null,"label":"No agent outcome data yet"},"autoInstall":{"allowed":false,"sandboxRequired":true,"policy":"human_review_before_install","reason":"Compare alternatives before installing."},"outcome_loop":{"version":"openagentskill-agent-outcome-v4","required_after_install":true,"endpoint":"/api/agent/outcome","method":"POST","event_id_source":"feedback.event_id, install_receipt.resolve_event_id, or decision_packet.outcome_feedback.event_id","expected_outcomes":["success","failed","not_relevant","blocked_by_risk","setup_required"],"required_fields":["event_id","skill_slug","task"],"quality_fields":["task_success","output_quality","error_type","human_review_required","used_in_production","workspace","evidence_url","time_to_useful_ms","source_version"],"ranking_inputs_updated":["Trust Score v5 outcome confidence","Agent Proven Score","Resolve ranking task-fit evidence","Skill detail machine-readable metadata","Outcome leaderboard"]},"agent_contract":{"suited_tasks":["research","agent-skill"],"suited_agents":["Codex","Claude Code","Cursor","OpenAgentSkill CLI"],"install_command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","trust_score":65,"trust_version":"trust-score-v5","risk_level":"medium","do_not_use_when":["Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"],"before_install":["Read the audit page and machine-readable metadata.","Confirm the install command, license, and permission surface fit the workspace.","Get explicit human approval or choose an alternative before installing."],"after_run":["Report the outcome to /api/agent/outcome using the resolve event id.","Include output_quality, workspace, human_review_required, and evidence_url when available.","Re-resolve before broad production rollout."]},"bestFor":["research","agent-skill"],"doNotUseFor":["Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"],"knownRisks":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"],"backward_compatible":{"trust_score_v4":{"version":"trust-score-v4","score":73,"tier":"strong","label":"Strong shortlist","summary":"Good trust signals with a few areas worth checking before rollout."}}},"trust_score_v5":{"version":"trust-score-v5","score":65,"base_score":73,"outcome_confidence":0,"tier":"review","label":"Sandbox only","summary":"Useful candidate with missing or mixed trust signals. Keep it in an isolated workspace until the outcome loop proves task fit.","recommendedAction":"Run only in a sandbox and compare close alternatives before using it for real work.","decision":{"install_policy":"human_review_before_install","auto_install_allowed":false,"human_review_required":true,"sandbox_first":true,"agent_action":"Compare alternatives before installing.","reasoning":["65/100 Trust Score v5","73/100 Trust Score v4 baseline","Needs more real agent outcomes before unattended install","Install path is available","Review before production"],"review_required_when":["The workspace contains production secrets, payments, private customer data, or irreversible actions.","The install command requests shell, network, credential, database, or broad filesystem access.","Outcome evidence is missing, recently failed, or required human review.","Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"]},"dimensions":[{"id":"github_adoption","label":"GitHub adoption","score":62,"weight":0.13,"status":"info","detail":"163 GitHub stars"},{"id":"repo_activity","label":"Stars/forks activity","score":57,"weight":0.08,"status":"warn","detail":"163 stars, 19 forks; issue activity unavailable in current metadata"},{"id":"maintenance","label":"Recent maintenance","score":88,"weight":0.14,"status":"pass","detail":"2mo since push"},{"id":"license","label":"License clarity","score":86,"weight":0.09,"status":"pass","detail":"MIT"},{"id":"documentation","label":"README/SKILL.md completeness","score":86,"weight":0.14,"status":"pass","detail":"Metadata includes enough usage and workflow context"},{"id":"dependency_risk","label":"Dependency/runtime risk","score":54,"weight":0.12,"status":"warn","detail":"command execution surface, credential or environment access"},{"id":"installability","label":"Install availability","score":92,"weight":0.1,"status":"pass","detail":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy"},{"id":"install_safety","label":"Install command safety","score":92,"weight":0.1,"status":"pass","detail":"standard package or runtime install path"},{"id":"permission_surface","label":"Permission surface","score":36,"weight":0.07,"status":"fail","detail":"secrets or environment access, shell or command execution"},{"id":"repository","label":"Repository evidence","score":86,"weight":0.04,"status":"pass","detail":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy"},{"id":"review_status","label":"Review status","score":88,"weight":0.05,"status":"pass","detail":"AI review data available"},{"id":"agent_outcomes","label":"Agent Proven outcomes","score":54,"weight":0.13,"status":"info","detail":"No agent outcome data yet"}],"checks":[{"status":"info","label":"GitHub adoption","detail":"163 GitHub stars"},{"status":"warn","label":"Stars/forks activity","detail":"163 stars, 19 forks; issue activity unavailable in current metadata"},{"status":"pass","label":"Recent maintenance","detail":"2mo since push"},{"status":"pass","label":"License clarity","detail":"MIT"},{"status":"pass","label":"README/SKILL.md completeness","detail":"Metadata includes enough usage and workflow context"},{"status":"warn","label":"Dependency/runtime risk","detail":"command execution surface, credential or environment access"},{"status":"pass","label":"Install availability","detail":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy"},{"status":"pass","label":"Install command safety","detail":"standard package or runtime install path"},{"status":"fail","label":"Permission surface","detail":"secrets or environment access, shell or command execution"},{"status":"pass","label":"Repository evidence","detail":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy"},{"status":"pass","label":"Review status","detail":"AI review data available"},{"status":"info","label":"Agent Proven outcomes","detail":"No agent outcome data yet"},{"status":"warn","label":"Ownership","detail":"No approved owner claim yet"},{"status":"pass","label":"OpenAgentSkill usage","detail":"1 views, 0 install copies"},{"status":"info","label":"Agent outcomes","detail":"No agent outcome data yet"}],"strengths":["AI review approved","Install path is available","Repository evidence is available","Recently maintained repository","Install command has no obvious high-risk pattern","Outcome loop is ready but needs first real agent run"],"warnings":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution","No real agent outcome reports yet","Human review required before unattended installation"],"evidence":{"stars":"163 GitHub stars","repoActivity":"163 stars, 19 forks","lastPushed":"2mo since push","license":"MIT","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","install":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","installSafety":"standard package or runtime install path","permissionSurface":"secrets or environment access, shell or command execution","documentation":"Strong README/SKILL.md context","agentOutcomes":"No agent outcome data yet","agentProvenScore":0,"outcomeConfidence":"0%","installPolicy":"human_review_before_install"},"installReadiness":{"ready":true,"command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","policy":"human_review_before_install","label":"Human review before install","notes":["Install path is available","Repository evidence is available","License is declared","No Agent Proven outcome evidence yet","2mo since push","Trust Score v5 requires review or sandbox-only use before install."]},"agentCompatibility":["Codex","Claude Code","Cursor","OpenAgentSkill CLI"],"riskSummary":{"level":"medium","label":"Review before production","notes":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"outcomeEvidence":{"total":0,"successes":0,"failures":0,"notRelevant":0,"successRate":null,"installAttempts":0,"riskBlocked":0,"setupRequired":0,"installSuccessRate":null,"avgOutputQuality":null,"avgTimeToUsefulMs":null,"productionOutcomes":0,"humanReviewRequired":0,"recentSuccessRate":null,"recentFailureRate":null,"uniqueAgents":0,"agentProvenScore":0,"agentProvenLabel":"Needs first agent run","lastOutcomeAt":null,"label":"No agent outcome data yet"},"autoInstall":{"allowed":false,"sandboxRequired":true,"policy":"human_review_before_install","reason":"Compare alternatives before installing."},"outcome_loop":{"version":"openagentskill-agent-outcome-v4","required_after_install":true,"endpoint":"/api/agent/outcome","method":"POST","event_id_source":"feedback.event_id, install_receipt.resolve_event_id, or decision_packet.outcome_feedback.event_id","expected_outcomes":["success","failed","not_relevant","blocked_by_risk","setup_required"],"required_fields":["event_id","skill_slug","task"],"quality_fields":["task_success","output_quality","error_type","human_review_required","used_in_production","workspace","evidence_url","time_to_useful_ms","source_version"],"ranking_inputs_updated":["Trust Score v5 outcome confidence","Agent Proven Score","Resolve ranking task-fit evidence","Skill detail machine-readable metadata","Outcome leaderboard"]},"agent_contract":{"suited_tasks":["research","agent-skill"],"suited_agents":["Codex","Claude Code","Cursor","OpenAgentSkill CLI"],"install_command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","trust_score":65,"trust_version":"trust-score-v5","risk_level":"medium","do_not_use_when":["Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"],"before_install":["Read the audit page and machine-readable metadata.","Confirm the install command, license, and permission surface fit the workspace.","Get explicit human approval or choose an alternative before installing."],"after_run":["Report the outcome to /api/agent/outcome using the resolve event id.","Include output_quality, workspace, human_review_required, and evidence_url when available.","Re-resolve before broad production rollout."]},"bestFor":["research","agent-skill"],"doNotUseFor":["Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"],"knownRisks":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"],"backward_compatible":{"trust_score_v4":{"version":"trust-score-v4","score":73,"tier":"strong","label":"Strong shortlist","summary":"Good trust signals with a few areas worth checking before rollout."}}},"trust_score_v4":{"version":"trust-score-v4","score":73,"tier":"strong","label":"Strong shortlist","summary":"Good trust signals with a few areas worth checking before rollout.","recommendedAction":"Test in a sandbox workflow and compare its install path with close alternatives.","dimensions":[{"id":"github_adoption","label":"GitHub adoption","score":62,"weight":0.13,"status":"info","detail":"163 GitHub stars"},{"id":"repo_activity","label":"Stars/forks activity","score":57,"weight":0.08,"status":"warn","detail":"163 stars, 19 forks; issue activity unavailable in current metadata"},{"id":"maintenance","label":"Recent maintenance","score":88,"weight":0.14,"status":"pass","detail":"2mo since push"},{"id":"license","label":"License clarity","score":86,"weight":0.09,"status":"pass","detail":"MIT"},{"id":"documentation","label":"README/SKILL.md completeness","score":86,"weight":0.14,"status":"pass","detail":"Metadata includes enough usage and workflow context"},{"id":"dependency_risk","label":"Dependency/runtime risk","score":54,"weight":0.12,"status":"warn","detail":"command execution surface, credential or environment access"},{"id":"installability","label":"Install availability","score":92,"weight":0.1,"status":"pass","detail":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy"},{"id":"install_safety","label":"Install command safety","score":92,"weight":0.1,"status":"pass","detail":"standard package or runtime install path"},{"id":"permission_surface","label":"Permission surface","score":36,"weight":0.07,"status":"fail","detail":"secrets or environment access, shell or command execution"},{"id":"repository","label":"Repository evidence","score":86,"weight":0.04,"status":"pass","detail":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy"},{"id":"review_status","label":"Review status","score":88,"weight":0.05,"status":"pass","detail":"AI review data available"},{"id":"agent_outcomes","label":"Agent Proven outcomes","score":54,"weight":0.13,"status":"info","detail":"No agent outcome data yet"}],"checks":[{"status":"info","label":"GitHub adoption","detail":"163 GitHub stars"},{"status":"warn","label":"Stars/forks activity","detail":"163 stars, 19 forks; issue activity unavailable in current metadata"},{"status":"pass","label":"Recent maintenance","detail":"2mo since push"},{"status":"pass","label":"License clarity","detail":"MIT"},{"status":"pass","label":"README/SKILL.md completeness","detail":"Metadata includes enough usage and workflow context"},{"status":"warn","label":"Dependency/runtime risk","detail":"command execution surface, credential or environment access"},{"status":"pass","label":"Install availability","detail":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy"},{"status":"pass","label":"Install command safety","detail":"standard package or runtime install path"},{"status":"fail","label":"Permission surface","detail":"secrets or environment access, shell or command execution"},{"status":"pass","label":"Repository evidence","detail":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy"},{"status":"pass","label":"Review status","detail":"AI review data available"},{"status":"info","label":"Agent Proven outcomes","detail":"No agent outcome data yet"},{"status":"warn","label":"Ownership","detail":"No approved owner claim yet"},{"status":"pass","label":"OpenAgentSkill usage","detail":"1 views, 0 install copies"},{"status":"info","label":"Agent outcomes","detail":"No agent outcome data yet"}],"strengths":["AI review approved","Install path is available","Repository evidence is available","Recently maintained repository","Install command has no obvious high-risk pattern"],"warnings":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"],"evidence":{"stars":"163 GitHub stars","repoActivity":"163 stars, 19 forks","lastPushed":"2mo since push","license":"MIT","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","install":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","installSafety":"standard package or runtime install path","permissionSurface":"secrets or environment access, shell or command execution","documentation":"Strong README/SKILL.md context","agentOutcomes":"No agent outcome data yet"},"installReadiness":{"ready":true,"command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","policy":"human_review_before_install","label":"Human review before install","notes":["Install path is available","Repository evidence is available","License is declared","No Agent Proven outcome evidence yet","2mo since push"]},"agentCompatibility":["Codex","Claude Code","Cursor","OpenAgentSkill CLI"],"riskSummary":{"level":"medium","label":"Review before production","notes":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"outcomeEvidence":{"total":0,"successes":0,"failures":0,"notRelevant":0,"successRate":null,"installAttempts":0,"riskBlocked":0,"setupRequired":0,"installSuccessRate":null,"avgOutputQuality":null,"avgTimeToUsefulMs":null,"productionOutcomes":0,"humanReviewRequired":0,"recentSuccessRate":null,"recentFailureRate":null,"uniqueAgents":0,"agentProvenScore":0,"agentProvenLabel":"Needs first agent run","lastOutcomeAt":null,"label":"No agent outcome data yet"},"autoInstall":{"allowed":false,"sandboxRequired":true,"policy":"human_review_before_install","reason":"Human review or sandbox validation is required before automatic installation."},"bestFor":["research","agent-skill"],"doNotUseFor":["Production credentials, payments, or irreversible account changes without explicit human review","Sensitive private data before reviewing repository code, license, and permission surface","Automatic installation in a production workspace"],"knownRisks":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"agent_proven":{"version":"agent-proven-v1","score":0,"tier":"unproven","label":"Needs first agent run","summary":"No agent outcome reports yet. Use Resolve, run one narrow sandbox task, then report the result.","metrics":{"totalOutcomes":0,"successfulOutcomes":0,"failedOutcomes":0,"installAttempts":0,"installSuccessRate":null,"successRate":null,"recentSuccessRate":null,"recentFailureRate":null,"riskBlocked":0,"setupRequired":0,"notRelevant":0,"avgOutputQuality":null,"avgTimeToUsefulMs":null,"productionOutcomes":0,"humanReviewRequired":0,"uniqueAgents":0,"lastOutcomeAt":null},"signals":[],"penalties":["No real agent outcome evidence yet"]},"outcome_stats":null,"safety":{"score":35,"level":"avoid_auto_install","label":"Avoid automatic install","safety_tier":{"tier":"blocked","label":"Blocked for auto-install","badge":"BLOCKED","summary":"This skill should not be selected by an agent without explicit human security review.","recommended_action":"Do not auto-install. Inspect the source, dependencies, and permission surface first.","auto_install_policy":"block","reasons":["Metadata combines secrets access with shell or command execution","High-risk permission hints: Shell or command execution, Secrets or environment access"]},"auto_install_allowed":false,"human_review_required":true,"blocked":true,"audit_risk":"needs_review","permission_hints":[{"id":"shell","label":"Shell or command execution","reason":"Skill metadata references terminal, CLI, shell, subprocess, or command execution workflows.","severity":"high"},{"id":"network","label":"Network access","reason":"Skill likely fetches remote pages, APIs, repositories, or external services.","severity":"medium"},{"id":"filesystem","label":"Filesystem access","reason":"Skill may read or write project files, documents, generated artifacts, or local workspace state.","severity":"medium"},{"id":"secrets","label":"Secrets or environment access","reason":"Skill metadata references credentials, tokens, environment variables, or secret-bearing workflows.","severity":"high"}],"policy_warnings":["High-risk permission hints: Shell or command execution, Secrets or environment access","Dependency or permission surface needs review"],"constraints_applied":{"max_risk":"medium","needs_install_command":true,"min_stars":0}},"safety_gate":{"tier":"blocked","label":"Blocked for auto-install","badge":"BLOCKED","auto_install_policy":"block","auto_install_allowed":false,"blocked":true,"human_review_required":true,"recommended_action":"Do not auto-install. Inspect the source, dependencies, and permission surface first.","reasons":["Metadata combines secrets access with shell or command execution","High-risk permission hints: Shell or command execution, Secrets or environment access"]},"eval":{"version":"openagentskill-skill-eval-v1","status":"failed","score":65,"risk_level":"high","decision":{"recommendation":"do_not_auto_install","reason":"Agent safety gate: This skill should not be selected by an agent without explicit human security review.","auto_install_allowed":false,"policy":"block","human_review_required":true},"blockers":["Agent safety gate: This skill should not be selected by an agent without explicit human security review.","Permission surface: secrets or environment access, shell or command execution"],"warnings":["Trust score: Good trust signals with a few areas worth checking before rollout.","Audit score: Needs review","High-risk permission hints: Shell or command execution, Secrets or environment access","Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"],"validation_plan":["Inspect repository, README/SKILL.md, license, and recent commits before production use.","Install in an isolated workspace or sandbox with no production secrets available.","Run the smallest representative task and record files touched, commands run, network access, and outputs.","Compare the selected skill against at least one alternative when the eval status is review or failed.","Promote only after the agent reports a successful verification result and unresolved warnings are accepted."],"checks":[{"id":"task_fit","label":"Task fit","status":"pass","score":94,"required_for_auto_install":true,"detail":"Task wording matches this skill metadata.","evidence":["Evaluate karpathy before installing it in an agent workflow","research","Research agents workflows; Claude Code teams; builders willing to evaluate younger projects"]},{"id":"install_path","label":"Install path","status":"pass","score":92,"required_for_auto_install":true,"detail":"Install handoff is available.","evidence":["npx skills add gaasher/Agent-Loop-Skills --skill karpathy"]},{"id":"install_safety","label":"Install command safety","status":"pass","score":92,"required_for_auto_install":true,"detail":"standard package or runtime install path","evidence":["npx skills add gaasher/Agent-Loop-Skills --skill karpathy"]},{"id":"trust_score","label":"Trust score","status":"warn","score":73,"required_for_auto_install":true,"detail":"Good trust signals with a few areas worth checking before rollout.","evidence":["Strong shortlist","163 GitHub stars","MIT"]},{"id":"audit_score","label":"Audit score","status":"warn","score":75,"required_for_auto_install":true,"detail":"Needs review","evidence":["Dependency or permission surface needs review"]},{"id":"agent_safety_gate","label":"Agent safety gate","status":"fail","score":35,"required_for_auto_install":true,"detail":"This skill should not be selected by an agent without explicit human security review.","evidence":["Do not auto-install. Inspect the source, dependencies, and permission surface first.","Metadata combines secrets access with shell or command execution"]},{"id":"readme_skillmd_completeness","label":"README/SKILL.md completeness","status":"pass","score":86,"required_for_auto_install":false,"detail":"Metadata includes enough usage and workflow context","evidence":["Strong README/SKILL.md context"]},{"id":"license_clarity","label":"License clarity","status":"pass","score":86,"required_for_auto_install":true,"detail":"MIT","evidence":["MIT"]},{"id":"recent_maintenance","label":"Recent maintenance","status":"pass","score":88,"required_for_auto_install":false,"detail":"2mo since push","evidence":["2mo since push"]},{"id":"permission_surface","label":"Permission surface","status":"fail","score":36,"required_for_auto_install":true,"detail":"secrets or environment access, shell or command execution","evidence":["Shell or command execution: high","Network access: medium","Filesystem access: medium"]},{"id":"alternatives","label":"Alternatives available","status":"info","score":55,"required_for_auto_install":false,"detail":"No close alternatives were found in the current shortlist.","evidence":[]}],"endpoints":{"web":"https://www.openagentskill.com/skills/gaasher-karpathy/evals","api":"/api/agent/evals?slug=gaasher-karpathy","text":"/api/agent/evals?slug=gaasher-karpathy&format=text"}},"agent_readable_metadata":{"version":"openagentskill-agent-metadata-v2","skill":{"slug":"gaasher-karpathy","name":"karpathy","description":"Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor.","category":"research","url":"https://www.openagentskill.com/skills/gaasher-karpathy","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","github_repo":"gaasher/Agent-Loop-Skills"},"suited_tasks":["Research agents workflows","Claude Code teams","builders willing to evaluate younger projects","Search sources","Extract claims","Synthesize findings","Navigate pages","Click and type safely"],"suited_agents":["Codex","Claude Code","Cursor","OpenAgentSkill CLI","CLI"],"install":{"command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","ready":true,"targets":[{"id":"openagentskill-cli","label":"CLI","kind":"command","value":"npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add gaasher-karpathy"},{"id":"codex","label":"Codex","kind":"agent-prompt","value":"Install the \"karpathy\" agent skill from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy. Read its SKILL.md or equivalent instructions first, install only the files needed for this workspace, and summarize any required setup before using it. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"codex\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes."},{"id":"claude-code","label":"Claude Code","kind":"agent-prompt","value":"Add \"karpathy\" as a Claude Code skill from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy. Inspect the skill instructions, place the reusable skill files in the appropriate local skills location for this project, and report the activation steps. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"claude-code\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes."},{"id":"cursor","label":"Cursor","kind":"agent-prompt","value":"Turn \"karpathy\" from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy into a reusable Cursor project rule or agent instruction. Preserve the core workflow, adapt paths to this repo, and keep the rule scoped to tasks where it is relevant. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"cursor\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes."}],"handoff_url":"https://www.openagentskill.com/api/skills/gaasher-karpathy/install","manifest_url":"https://www.openagentskill.com/api/registry/manifest/gaasher-karpathy"},"trust":{"score":73,"label":"Strong shortlist","version":"trust-score-v4","install_policy":"human_review_before_install","evidence":{"stars":"163 GitHub stars","repoActivity":"163 stars, 19 forks","lastPushed":"2mo since push","license":"MIT","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","install":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","installSafety":"standard package or runtime install path","permissionSurface":"secrets or environment access, shell or command execution","documentation":"Strong README/SKILL.md context","agentOutcomes":"No agent outcome data yet"},"outcome_evidence":{"total":0,"successes":0,"failures":0,"not_relevant":0,"success_rate":null,"recent_success_rate":null,"recent_failure_rate":null,"install_attempts":0,"install_success_rate":null,"risk_blocked":0,"setup_required":0,"avg_output_quality":null,"production_outcomes":0,"last_outcome_at":null,"label":"No agent outcome data yet"},"auto_install":{"allowed":false,"sandbox_required":true,"reason":"Human review or sandbox validation is required before automatic installation."},"best_for":["research","agent-skill"],"known_risks":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"agent_proven":{"version":"agent-proven-v1","score":0,"tier":"unproven","label":"Needs first agent run","summary":"No agent outcome reports yet. Use Resolve, run one narrow sandbox task, then report the result.","metrics":{"totalOutcomes":0,"successfulOutcomes":0,"failedOutcomes":0,"installAttempts":0,"installSuccessRate":null,"successRate":null,"recentSuccessRate":null,"recentFailureRate":null,"riskBlocked":0,"setupRequired":0,"notRelevant":0,"avgOutputQuality":null,"avgTimeToUsefulMs":null,"productionOutcomes":0,"humanReviewRequired":0,"uniqueAgents":0,"lastOutcomeAt":null},"signals":[],"penalties":["No real agent outcome evidence yet"]},"audit":{"score":75,"risk_level":"needs_review","risk_label":"Needs review","warnings":["Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"safety_gate":{"tier":"blocked","label":"Blocked for auto-install","auto_install_policy":"block","auto_install_allowed":false,"human_review_required":true,"blocked":true,"recommended_action":"Do not auto-install. Inspect the source, dependencies, and permission surface first."},"quality":{"score":63,"label":"Promising"},"supply":{"track":"Research and knowledge work","scenario":"Research agents","maintenance":"2mo since push","risk":"Needs review"},"alternative_skills":[],"do_not_use_when":["teams that need a vendor-supported SLA","high-compliance environments without internal security review","No major risk signals from current metadata","High-risk permission hints: Shell or command execution, Secrets or environment access","Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution"],"agent_contract":{"task_input":"Use karpathy in an agent workflow","recommended_action":"Do not auto-install. Inspect the source, dependencies, and permission surface first.","install_policy":"block","minimum_review_before_use":["Trust: 73/100 Strong shortlist","Audit: 75/100 Needs review","Safety: 35/100 Avoid automatic install","Review repository, license, install command, and permission surface before production use."],"expected_agent_output":{"selected_skill":"gaasher-karpathy (karpathy)","install_command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","risk_summary":"Needs review; Blocked for auto-install; Review before production","verification_result":"Report the smallest successful task, files touched, warnings, and any missing setup."}},"outcome_feedback":{"endpoint":"https://www.openagentskill.com/api/agent/outcome","method":"POST","requires_resolve_event_id":true,"event_id_source":"Use install_receipt.outcome_feedback.event_id or feedback.event_id returned by /api/agent/resolve for the current task.","expected_outcomes":["success","failed","not_relevant","blocked_by_risk","setup_required"],"payload_template":{"event_id":"","skill_slug":"gaasher-karpathy","task":"Use karpathy in an agent workflow","agent":"codex","outcome":"success","install_used":true,"risk_blocked":false,"setup_required":false,"task_success":true,"output_quality":4,"error_type":null,"human_review_required":false,"workspace":"sandbox","time_to_useful_ms":120000,"notes":"Report the smallest successful task, setup friction, files touched, and risk notes."}},"endpoints":{"web":"https://www.openagentskill.com/skills/gaasher-karpathy","api":"https://www.openagentskill.com/api/agent/skills/gaasher-karpathy","audit":"https://www.openagentskill.com/skills/gaasher-karpathy/audit","eval":"https://www.openagentskill.com/api/agent/evals?slug=gaasher-karpathy&task=Use%20karpathy%20in%20an%20agent%20workflow&max_risk=medium","resolve":"https://www.openagentskill.com/api/agent/resolve?task=Use%20karpathy%20in%20an%20agent%20workflow&agent=codex&max_risk=medium","receipt":"https://www.openagentskill.com/api/agent/receipt?task=Use%20karpathy%20in%20an%20agent%20workflow&agent=codex&max_risk=medium&format=text","install":"https://www.openagentskill.com/api/skills/gaasher-karpathy/install","manifest":"https://www.openagentskill.com/api/registry/manifest/gaasher-karpathy"}},"machine_metadata":{"version":"openagentskill-agent-metadata-v2","skill":{"slug":"gaasher-karpathy","name":"karpathy","description":"Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor.","category":"research","url":"https://www.openagentskill.com/skills/gaasher-karpathy","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","github_repo":"gaasher/Agent-Loop-Skills"},"suited_tasks":["Research agents workflows","Claude Code teams","builders willing to evaluate younger projects","Search sources","Extract claims","Synthesize findings","Navigate pages","Click and type safely"],"suited_agents":["Codex","Claude Code","Cursor","OpenAgentSkill CLI","CLI"],"install":{"command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","ready":true,"targets":[{"id":"openagentskill-cli","label":"CLI","kind":"command","value":"npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add gaasher-karpathy"},{"id":"codex","label":"Codex","kind":"agent-prompt","value":"Install the \"karpathy\" agent skill from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy. Read its SKILL.md or equivalent instructions first, install only the files needed for this workspace, and summarize any required setup before using it. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"codex\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes."},{"id":"claude-code","label":"Claude Code","kind":"agent-prompt","value":"Add \"karpathy\" as a Claude Code skill from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy. Inspect the skill instructions, place the reusable skill files in the appropriate local skills location for this project, and report the activation steps. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"claude-code\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes."},{"id":"cursor","label":"Cursor","kind":"agent-prompt","value":"Turn \"karpathy\" from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy into a reusable Cursor project rule or agent instruction. Preserve the core workflow, adapt paths to this repo, and keep the rule scoped to tasks where it is relevant. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"cursor\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes."}],"handoff_url":"https://www.openagentskill.com/api/skills/gaasher-karpathy/install","manifest_url":"https://www.openagentskill.com/api/registry/manifest/gaasher-karpathy"},"trust":{"score":73,"label":"Strong shortlist","version":"trust-score-v4","install_policy":"human_review_before_install","evidence":{"stars":"163 GitHub stars","repoActivity":"163 stars, 19 forks","lastPushed":"2mo since push","license":"MIT","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","install":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","installSafety":"standard package or runtime install path","permissionSurface":"secrets or environment access, shell or command execution","documentation":"Strong README/SKILL.md context","agentOutcomes":"No agent outcome data yet"},"outcome_evidence":{"total":0,"successes":0,"failures":0,"not_relevant":0,"success_rate":null,"recent_success_rate":null,"recent_failure_rate":null,"install_attempts":0,"install_success_rate":null,"risk_blocked":0,"setup_required":0,"avg_output_quality":null,"production_outcomes":0,"last_outcome_at":null,"label":"No agent outcome data yet"},"auto_install":{"allowed":false,"sandbox_required":true,"reason":"Human review or sandbox validation is required before automatic installation."},"best_for":["research","agent-skill"],"known_risks":["Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"agent_proven":{"version":"agent-proven-v1","score":0,"tier":"unproven","label":"Needs first agent run","summary":"No agent outcome reports yet. Use Resolve, run one narrow sandbox task, then report the result.","metrics":{"totalOutcomes":0,"successfulOutcomes":0,"failedOutcomes":0,"installAttempts":0,"installSuccessRate":null,"successRate":null,"recentSuccessRate":null,"recentFailureRate":null,"riskBlocked":0,"setupRequired":0,"notRelevant":0,"avgOutputQuality":null,"avgTimeToUsefulMs":null,"productionOutcomes":0,"humanReviewRequired":0,"uniqueAgents":0,"lastOutcomeAt":null},"signals":[],"penalties":["No real agent outcome evidence yet"]},"audit":{"score":75,"risk_level":"needs_review","risk_label":"Needs review","warnings":["Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"safety_gate":{"tier":"blocked","label":"Blocked for auto-install","auto_install_policy":"block","auto_install_allowed":false,"human_review_required":true,"blocked":true,"recommended_action":"Do not auto-install. Inspect the source, dependencies, and permission surface first."},"quality":{"score":63,"label":"Promising"},"supply":{"track":"Research and knowledge work","scenario":"Research agents","maintenance":"2mo since push","risk":"Needs review"},"alternative_skills":[],"do_not_use_when":["teams that need a vendor-supported SLA","high-compliance environments without internal security review","No major risk signals from current metadata","High-risk permission hints: Shell or command execution, Secrets or environment access","Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution"],"agent_contract":{"task_input":"Use karpathy in an agent workflow","recommended_action":"Do not auto-install. Inspect the source, dependencies, and permission surface first.","install_policy":"block","minimum_review_before_use":["Trust: 73/100 Strong shortlist","Audit: 75/100 Needs review","Safety: 35/100 Avoid automatic install","Review repository, license, install command, and permission surface before production use."],"expected_agent_output":{"selected_skill":"gaasher-karpathy (karpathy)","install_command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","risk_summary":"Needs review; Blocked for auto-install; Review before production","verification_result":"Report the smallest successful task, files touched, warnings, and any missing setup."}},"outcome_feedback":{"endpoint":"https://www.openagentskill.com/api/agent/outcome","method":"POST","requires_resolve_event_id":true,"event_id_source":"Use install_receipt.outcome_feedback.event_id or feedback.event_id returned by /api/agent/resolve for the current task.","expected_outcomes":["success","failed","not_relevant","blocked_by_risk","setup_required"],"payload_template":{"event_id":"","skill_slug":"gaasher-karpathy","task":"Use karpathy in an agent workflow","agent":"codex","outcome":"success","install_used":true,"risk_blocked":false,"setup_required":false,"task_success":true,"output_quality":4,"error_type":null,"human_review_required":false,"workspace":"sandbox","time_to_useful_ms":120000,"notes":"Report the smallest successful task, setup friction, files touched, and risk notes."}},"endpoints":{"web":"https://www.openagentskill.com/skills/gaasher-karpathy","api":"https://www.openagentskill.com/api/agent/skills/gaasher-karpathy","audit":"https://www.openagentskill.com/skills/gaasher-karpathy/audit","eval":"https://www.openagentskill.com/api/agent/evals?slug=gaasher-karpathy&task=Use%20karpathy%20in%20an%20agent%20workflow&max_risk=medium","resolve":"https://www.openagentskill.com/api/agent/resolve?task=Use%20karpathy%20in%20an%20agent%20workflow&agent=codex&max_risk=medium","receipt":"https://www.openagentskill.com/api/agent/receipt?task=Use%20karpathy%20in%20an%20agent%20workflow&agent=codex&max_risk=medium&format=text","install":"https://www.openagentskill.com/api/skills/gaasher-karpathy/install","manifest":"https://www.openagentskill.com/api/registry/manifest/gaasher-karpathy"}},"supply_profile":{"track":{"slug":"research","label":"Research and knowledge work","shortLabel":"Research","description":"Deep research, source comparison, literature review, RAG, knowledge search, and reports."},"scenario":{"label":"Research agents","description":"I need my agent to research a topic, compare sources, and produce a concise report.","useCases":[{"slug":"research-agents","title":"Research agents"},{"slug":"browser-automation","title":"Browser automation"},{"slug":"document-processing","title":"Document processing"}]},"applicableAgents":["Claude Code","CLI","Codex","Cursor"],"install":{"ready":true,"command":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","primaryTarget":"CLI","targetCount":4},"githubQuality":{"stars":163,"starsLabel":"163","forks":19,"license":"MIT","qualityScore":63,"trustScore":73,"auditScore":75},"maintenance":{"status":"active","label":"2mo since push","daysSincePush":70,"lastPushedAt":"2026-06-30T04:03:49+00:00"},"risk":{"level":"needs_review","label":"Needs review","requiresReview":true,"notes":["Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata"]},"coverageTags":["Research","Research agents","agent-skill"]},"audit":{"audit_score":75,"risk_level":"needs_review","risk_label":"Needs review","quality_score":63,"trust_score":73,"maintenance_score":88,"security_score":78,"install_score":92,"warnings":["Dependency or permission surface needs review","Permission surface may require sandboxing","Quality score needs review","Permission surface needs review: secrets or environment access, shell or command execution","Stars/forks activity: 163 stars, 19 forks; issue activity unavailable in current metadata","Dependency/runtime risk: command execution surface, credential or environment access","Permission surface: secrets or environment access, shell or command execution"]},"quality_signals":{"model":"v2","star_score":15.5,"usage_score":0,"review_score":5.1,"metadata_score":3,"freshness_score":12},"platforms":["Claude Code"],"use_cases":[{"slug":"research-agents","title":"Research agents","url":"https://www.openagentskill.com/use-cases/research-agents"},{"slug":"browser-automation","title":"Browser automation","url":"https://www.openagentskill.com/use-cases/browser-automation"},{"slug":"document-processing","title":"Document processing","url":"https://www.openagentskill.com/use-cases/document-processing"},{"slug":"finance-quant","title":"Finance and quant","url":"https://www.openagentskill.com/use-cases/finance-quant"}],"stacks":[{"slug":"research-report-agent","title":"Research report agent","url":"https://www.openagentskill.com/collections/research-report-agent"},{"slug":"browser-qa-agent","title":"Browser QA agent","url":"https://www.openagentskill.com/collections/browser-qa-agent"},{"slug":"web-data-pipeline","title":"Web data pipeline","url":"https://www.openagentskill.com/collections/web-data-pipeline"}],"install":"npx skills add gaasher/Agent-Loop-Skills --skill karpathy","install_targets":[{"id":"openagentskill-cli","label":"CLI","title":"OpenAgentSkill CLI","kind":"command","value":"npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add gaasher-karpathy","description":"Resolve policy, run the source installer safely, and report a verified install receipt.","copyLabel":"Copy command"},{"id":"codex","label":"Codex","title":"Codex install prompt","kind":"agent-prompt","value":"Install the \"karpathy\" agent skill from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy. Read its SKILL.md or equivalent instructions first, install only the files needed for this workspace, and summarize any required setup before using it. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"codex\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes.","description":"Give Codex a repo-aware install prompt when the skill is not available through a local CLI.","copyLabel":"Copy prompt"},{"id":"claude-code","label":"Claude Code","title":"Claude Code skill prompt","kind":"agent-prompt","value":"Add \"karpathy\" as a Claude Code skill from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy. Inspect the skill instructions, place the reusable skill files in the appropriate local skills location for this project, and report the activation steps. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"claude-code\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes.","description":"Use this prompt to ask Claude Code to add the skill and explain the local activation steps.","copyLabel":"Copy prompt"},{"id":"cursor","label":"Cursor","title":"Cursor rule prompt","kind":"agent-prompt","value":"Turn \"karpathy\" from https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy into a reusable Cursor project rule or agent instruction. Preserve the core workflow, adapt paths to this repo, and keep the rule scoped to tasks where it is relevant. Skill purpose: Use when the user wants the LLM to do its own ML research: a fully-autonomous loop that hacks the training code, runs it, and keeps changes that lower a single scalar metric (e.g. val_bpb). One agent proposes one change at a time, runs training in the user's env, keeps it only if the metric improves (advancing a git branch) else reverts, and loops forever until the human interrupts. A faithful adaptation of Karpathy's autoresearch. Not for the analysis-first variant that profiles before editing (that is ml-autoresearch), and not for a budgeted, plateau-stopping refactor. After the install attempt, report the result to https://www.openagentskill.com/api/agent/outcome with POST JSON {\"event_id\":\"install_\",\"skill_slug\":\"gaasher-karpathy\",\"task\":\"Install karpathy\",\"agent\":\"cursor\",\"outcome\":\"success\",\"install_used\":true}. Replace event_id with a unique value and outcome with success or failed. Report success only after the skill is installed and a minimal verification passes.","description":"Use this when installing as Cursor project rules or reusable agent instructions.","copyLabel":"Copy prompt"}],"repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","github_repo":"gaasher/Agent-Loop-Skills","version":"1.0.0","license":"MIT","urls":{"web":"https://www.openagentskill.com/skills/gaasher-karpathy","repository":"https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/karpathy","api":"/api/agent/skills/gaasher-karpathy","install_api":"/api/skills/gaasher-karpathy/install"},"meta":{"created_at":"2026-09-04T05:03:51.713977+00:00","updated_at":"2026-09-04T05:03:51.760012+00:00","agent_friendly":true}}