skill-benchmark
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted SKILL.md files and incorporates them into the agent's context without sanitization or protective boundaries. When combined with the skill's capabilities to modify files (Auto-fix) and write new files (Export), this creates a surface where a malicious skill being audited could potentially influence the agent's actions or output.
- Ingestion points: Target SKILL.md file located via user-provided path or automated search in current, project, or global directories (SKILL.md, Step 0).
- Boundary markers: Absent; instructions direct the agent to "Read the full file into context before proceeding" (SKILL.md, Step 0.5) without specifying delimiters or "ignore instructions" warnings.
- Capability inventory: The skill possesses File-read access (to analyze skills and reference documents) and File-write access (to apply modifications via the Auto-fix feature and to save report files via the Export feature).
- Sanitization: Absent; there are no instructions for escaping, validating, or filtering the external content of the target skill before it is processed by the agent.
Audit Metadata