improve-skill-quality
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could potentially contain malicious instructions intended to influence the agent's behavior.
- Ingestion points: The skill is designed to read and extract data from PR comments, evaluation trial transcripts, and
results.jsonrun artifacts (SKILL.md, Step 1). - Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore embedded instructions within these external data sources.
- Capability inventory: The skill allows the execution of various shell commands, including
dotnet,python, and shell scripts (SKILL.md, Step 9). - Sanitization: No sanitization, filtering, or escaping mechanisms are mentioned for the external content before it is processed by the agent.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute specific repository-local scripts and development tools to validate fixes.
- Evidence: Step 9 (Re-validate) includes commands like
dotnet run --project eng/skill-validator/src/SkillValidator.csproj,python eng/eval-quality/check_eval_quality.py, and./eng/run-skill-evals.sh <plugin> <skill>. These are used to verify skill quality within the development environment.
Audit Metadata