grant-writer

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill accesses and writes to the path ~/.claude/projects/ to record execution metadata. Accessing platform-specific hidden directories for logging purposes without explicit user disclosure in the metadata represents a minor concern regarding filesystem privacy.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. Ingestion points: the entire codebase via file read operations (Phases 1-8). Boundary markers: Absent. Capability inventory: Extensive filesystem read access and writing to local telemetry logs. Sanitization: Absent. Malicious instructions within the audited codebase could potentially influence the agent's behavior or output.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:57 AM
Security Audit — agent-trust-hub — grant-writer