dead-code-detector

Pass

Audited by Gen Agent Trust Hub on Oct 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONPERSISTENCE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes untrusted data from external codebases, including source code, comments (TODOs/FIXMEs), and documentation files, which could contain malicious instructions intended to influence the agent's validation logic.
  • Ingestion points: Source code, READMEs, design documents, and issue trackers are read by the 'Intent', 'Integration', and 'History' subagents defined in SKILL.md.
  • Boundary markers: The prompts provided for the subagents do not include specific delimiters or instructions to ignore potential commands embedded in the analyzed text.
  • Capability inventory: The skill has access to Bash, Edit, and Write tools, enabling it to delete code or modify files based on its analysis.
  • Sanitization: No sanitization or escaping of the external content is performed before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to install dependencies and execute analysis tools.
  • Evidence: Commands such as uv pip install vulture, bun add -d knip, and cargo clippy --fix --allow-dirty are used to manage the environment and perform code modifications.
  • [PERSISTENCE]: The skill contains explicit instructions for the agent to modify the skill's own source file (SKILL.md) to reflect improvements or workarounds discovered during execution.
  • Evidence: The 'Self-Evolving Skill' and 'Post-Execution Reflection' sections in SKILL.md direct the agent to "fix this file immediately" and "update this SKILL.md so the next invocation doesn't need the same workaround."
  • Risk: This self-modification pattern allows for the permanent alteration of the skill's behavior, which could be abused to persist malicious instructions if the agent is compromised during a session.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 17, 2026, 06:24 AM