skills/coakenfold/p-100226/debug/Gen Agent Trust Hub

debug

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted input from error messages and source code while possessing command execution capabilities.
  • Ingestion points: The skill processes user-provided error descriptions ($ARGUMENTS) and reads files from the codebase (using Read, Glob, and Grep tools) in Phase 1 and Phase 3.
  • Boundary markers: There are no explicit instructions or delimiters used to separate data from instructions or to warn the agent to ignore instructions embedded in the analyzed code or stack traces.
  • Capability inventory: The skill utilizes powerful tools including Bash, Edit, Write, and Task, allowing it to modify files and execute arbitrary shell commands like 'npm test'.
  • Sanitization: No sanitization or validation of the ingested content is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 01:21 AM
Security Audit — agent-trust-hub — debug