systematic-debugging
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to search for and read content from external sources including official documentation, GitHub Issues, and Stack Overflow. This content is untrusted and represents a surface for indirect prompt injection where malicious instructions could be embedded in the data the agent processes.\n
- Ingestion points: Data retrieved from external documentation, GitHub issue trackers, and web search results as specified in
references/external-research-guide.md.\n - Capability inventory: The agent is authorized to modify source code, execute git commands, and run test suites via
vitestto verify fixes as outlined inSKILL.md.\n - Boundary markers: The skill emphasizes careful reading and classification of symptoms but does not specify technical delimiters or explicit instructions to ignore prompt-like content within the fetched data.\n
- Sanitization: No automated filtering or sanitization of external research results is mentioned.\n- [COMMAND_EXECUTION]: The skill workflow involves executing command-line tools for environment inspection, testing, and version control.\n
- Evidence: Mentions of
git diff,git checkout,git stash, andnpx vitestfor identifying changes, managing worktree state, and verifying fixes as defined inSKILL.mdandreferences/root-cause-tracing.md.
Audit Metadata