extract-requirements
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied legacy codebases which act as an untrusted data source. Instructions or malicious patterns embedded within code comments or string literals could potentially be interpreted by the AI agent during the extraction and analysis phases.\n
- Ingestion points: The skill ingests codebases via the
codecompass batch:index <path>command and subsequent analysis steps as described inSKILL.md.\n - Boundary markers: There are no specific instructions or delimiters provided to the agent to treat the analyzed code strictly as data and ignore any embedded instructions.\n
- Capability inventory: The agent utilizes the
codecompassCLI for extraction and semantic search, and has the ability to write output to local files likerequirements.md.\n - Sanitization: No sanitization or filtering of the input codebase is performed before processing.\n- [COMMAND_EXECUTION]: The skill relies on the execution of the
codecompassCLI tool to perform its primary functions.\n - Evidence: Use of commands such as
codecompass batch:index <path>,codecompass requirements:extract --project-id <id> --output requirements.md, andcodecompass search:semanticthroughout the workflow.
Audit Metadata