init-context

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests and processes untrusted data from the repository's history, including documentation, PR/issue discussions, and code comments. This creates a surface for indirect prompt injection where malicious content in the project history could influence the agent's reasoning or documentation output.
  • Ingestion points: As described in references/MINING.md, the skill reads prose documentation, PR/Issue discussions, source code, and commit history.
  • Boundary markers: No explicit delimiters or boundary markers are utilized to isolate the ingested history data from the agent's core instructions.
  • Capability inventory: The skill possesses the capability to perform local file system writes, specifically creating and modifying docs/adr/CONVENTIONS.md, CONTEXT.md, and project context files like CLAUDE.md.
  • Sanitization: The skill mitigates risks through a 'Ratify, don't write' policy for vocabulary entries, ensuring that generated glossary candidates are presented to the user for approval rather than being committed automatically.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 09:47 PM
Security Audit — agent-trust-hub — init-context