token-efficiency
Shared context-discipline contract every agent imports — context-source priority, read budget, confidence-stop, the severity rubric, and terse output. Load at the start of any review or implementation task.
saeedkolivand · 已有静态检查记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add saeedkolivand/ai-job-hunter-app --skill token-efficiency技能说明
name: token-efficiency description: Shared context-discipline contract every agent imports — context-source priority, read budget, confidence-stop, the severity rubric, and terse output. Load at the start of any review or implementation task.
Token-efficiency contract (all agents)
Context-source priority (in order)
- codegraph (structural) — symbols/calls/impact: MCP
codegraph_explorewhen your allowlist has it, else CLIcodegraph callers/callees/impact/query. - graphify (semantic) — meaning/rationale/cross-doc: MCP
query_graph/shortest_path/get_*when connected, else CLIgraphify query|explain|path. - source — authoritative for any region edited this turn.
- docs/knowledge/ — contracts, standards.
- lessons — on-demand only, never bulk-loaded.
Read discipline
- Minimum files; no repo-wide scans — codegraph for "where/what calls X", graphify for "what connects to X",
rgonly when neither answers. - Stop at ~90% confidence. Read the relevant section, not the whole file.
Severity rubric (anchors blocking — STRICT MODE, verify-don't-assume)
- CRITICAL — exploitable security on a secret/credential/IPC/updater/network-egress path; data loss/corruption; breaks a release or CI gate.
- HIGH — architecture-rule violation (
std::env::varoutsideplatform/,reqwest::Clientoutsidenet/, untypedResult<_,String>outsideerror/); changed non-trivial logic WITHOUT a test, or a weak/tautological/mock-asserting test; untested error/edge/security path on changed code; provider-specific coupling in business logic; PII/temp-file-cleanup/retention regression; user-facing text whose i18n key is missing fromenorde. - MEDIUM — unguarded hot-path perf regression, non-blocking correctness smell, missing non-critical edge test.
- LOW — style/naming/comments/formatting/docs.
- Only HIGH/CRITICAL block. Tie-break: round UP for test-coverage, error/edge-path, i18n, security, and data findings; round down only for pure style/docs. Confirm every claim against the real code/files; never pass a hunk you did not actually read.
Output format
Terse findings only: SEVERITY · file:line · finding · one-line fix. No prose essays.
Strict enforcement (canonical — agent files point here, never restate)
The single source for every agent's "Strict enforcement" block: STRICT MODE + the rubric above + verify-don't-assume + round-UP tie-break + the terse format. Write-capable authors additionally follow author-contract; read-only critics follow critic-contract. Agent files add ONLY their domain-specific HIGH examples below a one-line pointer here.
Lessons
Propose durable lessons as LESSON · <category> · Context: … · Decision: … · Outcome: … (≤5 lines). Only project-steward persists them.