pr-babysit Eval
===============

Status: failed
Score: 63/100
Risk: high
Decision: do_not_auto_install
Policy: block
Reason: Permission surface: shell or command execution, filesystem or document access

Install:
npx skills add VincentChuWaiChow/vanguard-frontier-agentic --skill pr-babysit

Required checks:
- PASS Task fit: Task wording matches this skill metadata.
- PASS Install path: Install handoff is available.
- PASS Install command safety: standard package or runtime install path
- WARN Trust score: Potentially useful, but at least one trust signal needs human inspection.
- WARN Audit score: Needs review
- WARN Agent safety gate: Sparse or mixed signals. Useful for discovery, but not for autonomous installation.
- PASS License clarity: Apache-2.0
- FAIL Permission surface: shell or command execution, filesystem or document access

Warnings:
- Trust score: Potentially useful, but at least one trust signal needs human inspection.
- Audit score: Needs review
- Agent safety gate: Sparse or mixed signals. Useful for discovery, but not for autonomous installation.
- README/SKILL.md completeness: Public metadata needs stronger README/SKILL.md context
- High-risk permission hints: Shell or command execution
- Permission surface may require sandboxing
- allowed-tools includes unrestricted Bash without explicit guardrails for destructive or network-changing commands; a malicious CI log or PR comment could potentially influence command construction.
- SKILL.md references multiple external mechanisms and skill dependencies that are not fully specified: the harness subscribe tool, scheduling primitive, GitHub MCP pull_request_read, agentic-delegation, and definition-of-done.
- allowed-tools does not include the GitHub MCP or subscribe/scheduling tools referenced in the workflow, creating a mismatch between declared tools and required tools.
- Low GitHub adoption signal
- Quality score needs review
- Permission surface needs review: shell or command execution, filesystem or document access

Validation plan:
1. Inspect repository, README/SKILL.md, license, and recent commits before production use.
2. Install in an isolated workspace or sandbox with no production secrets available.
3. Run the smallest representative task and record files touched, commands run, network access, and outputs.
4. Compare the selected skill against at least one alternative when the eval status is review or failed.
5. Promote only after the agent reports a successful verification result and unresolved warnings are accepted.

Do not use when:
- teams that need a vendor-supported SLA
- production agents without a repository review
- Low GitHub adoption signal
- allowed-tools includes unrestricted Bash without explicit guardrails for destructive or network-changing commands; a malicious CI log or PR comment could potentially influence command construction.
- No OpenAgentSkill engagement data yet
- High-risk permission hints: Shell or command execution
- Permission surface may require sandboxing
- SKILL.md references multiple external mechanisms and skill dependencies that are not fully specified: the harness subscribe tool, scheduling primitive, GitHub MCP pull_request_read, agentic-delegation, and definition-of-done.

URLs:
- Skill: https://www.openagentskill.com/skills/vincentchuwaichow-pr-babysit
- Audit: https://www.openagentskill.com/skills/vincentchuwaichow-pr-babysit/audit
- JSON: https://www.openagentskill.com/api/agent/evals?slug=vincentchuwaichow-pr-babysit