Security audit
Security checks for vulnerabilities and agentic risk
This markdown-only skill is a disclosed completion-audit workflow with no code execution, persistence, credential use, or data exfiltration, though its auto-invocation behavior deserves user awareness.
Install this if you want an agent to proactively audit completed work for fake or shallow completion. Be aware it is designed to run without a separate explicit request when it sees a completion claim and can recover the original purpose, so it may inspect surrounding task or repository context and produce an audit at handoff time.
The skill is configured to auto-invoke on common completion language such as 'done' or 'all tests pass' whenever it can infer prior purpose from context. That trigger condition is broad enough to fire during ordinary workflow handoffs, causing unrequested auditing, context overreach, and possible disruption or leakage of unrelated task context into the audit.
This section explicitly instructs the skill to auto-trigger without permission based on ambiguous completion indicators and inferred context. In an agentic system, that creates a policy-level safety risk: the skill may activate on benign completion chatter, perform unsolicited analysis, and override user expectations about when specialized auditing should occur.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
---
name: impostor-hunt
description: |
Hunt the impostor in a "finished" deliverable. An impostor is a result-correct artifact whose causal chain is NOT the one the user's purpose required — the test set leaked, only the happy path is wired, correlation got relabeled as cause, the config file exists but no code path reads it, the wrapper greps stdout for "OK" without checking the exit code, the agent "can refactor" backed by one cherry-picked example. The output looks right. The logic that produced "right" is the wrong logic.
AUTO-INVOKE when a turn declares completion AND the user's original purpose is recoverable from context. Completion signals include: "done", "implemented", "finished", "works now", "ready to merge", "all tests pass", "task complete", "feature shipped", an agent posting a summary with checkmarks, a PR description, a CHANGELOG entry, or a closing message that asserts the work is over. Original purpose is recoverable from: the user's first message in the thread, the issue/ticket text, the PR title and description, the commit message that opened the branch, the README's stated goal, or an explicit prompt the user pasted at the start. If both signals are present, invoke without waiting for an explicit request.
No suspicious patterns detected.