context-audit
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface identified. The skill ingests untrusted data from AI context files (e.g., CLAUDE.md, .cursorrules) and repository configuration files which are processed by subagents to perform quality analysis. Malicious instructions embedded in these audited files could potentially influence agent behavior during the audit or remediation phases. Ingestion points: AI context files and repository manifests (package.json, Cargo.toml, etc.) as defined in the discovery workflow. Boundary markers: No specific delimiters or instruction isolation techniques are documented to separate audited content from the analysis prompts. Capability inventory: The skill has file-write permissions for resolving identified issues and generating new context templates. Sanitization: No explicit sanitization or filtering of external file content is performed before processing.
- [COMMAND_EXECUTION]: The skill utilizes repository search tools including Glob and Grep to validate the existence of file paths, directory references, and code symbols mentioned in documentation.
Audit Metadata