agentsmd
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to identify and run commands found within the repository's build files, scripts, and CI configurations to confirm they work.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from untrusted repository files to generate instruction content, which could allow malicious instructions present in those files to be incorporated into the agent's persistent instruction set. 1. Ingestion points: README.md, build files, repository scripts, CI configurations, and existing CLAUDE.md files. 2. Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded prompts when processing repo data. 3. Capability inventory: The agent has the ability to read arbitrary files in the repository, write to AGENTS.md and CLAUDE.md, and execute shell commands discovered in build or script files. 4. Sanitization: No explicit sanitization or filtering of the ingested content is performed before it is written to the instruction files.
Audit Metadata