context-health

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses basic shell utilities (wc -l, head, grep) to analyze the metadata and contents of local markdown files in the ~/.claude/ directory. These operations are read-only and used for the stated audit purpose.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes content from local context files that could contain embedded instructions.
  • Ingestion points: Content is read from multiple markdown files in ~/.claude/.context/core/, including identity.md, preferences.md, and rules.md (SKILL.md).
  • Boundary markers: Absent. The skill does not implement delimiters or 'ignore' instructions to prevent the agent from accidentally obeying commands found within the audited context data.
  • Capability inventory: The skill is limited to reading local files and executing standard text-processing shell commands; it does not exhibit network exfiltration or destructive file-write capabilities (SKILL.md).
  • Sanitization: No escaping, validation, or filtering is performed on the content read from the local files before it is presented to the agent for analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 01:57 AM