skills/ferueda/harness/create-plan/Gen Agent Trust Hub

create-plan

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by processing untrusted data to create plan files.\n
  • Ingestion points: Processes external inputs such as 'todo, spec, issue, review notes, or raw instructions' as specified in SKILL.md.\n
  • Boundary markers: Lacks explicit boundary markers or delimiters for the untrusted input within the prompt instructions to distinguish user data from agent instructions.\n
  • Capability inventory: The agent possesses capabilities to read source code, write files to the dev/plans/ directory, and execute repository verification commands.\n
  • Sanitization: No explicit sanitization or validation of the input content is defined in the skill instructions to prevent instruction override.\n- [SAFE]: The skill enforces a quality bar in references/plan-template.md that explicitly prohibits the inclusion of secrets in generated plans, which reduces the risk of accidental data exposure.\n- [SAFE]: No evidence of obfuscation, malicious external downloads, credential theft, or unauthorized persistence mechanisms was found in the skill's instructions or configuration files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 09:38 PM
Security Audit — agent-trust-hub — create-plan