skill-improver

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the plugin-dev plugin from the Trail of Bits repository. Trail of Bits is a well-known security research firm.
  • [PROMPT_INJECTION]: The skill uses behavioral override instructions, stating that the loop MUST ONLY be terminated by a specific marker ('') and that natural language completion signals should be disregarded.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external files at a user-provided path, which introduces a surface where instructions within those files could influence the agent's actions during the fix-review cycle. 1. Ingestion points: Skill files located at [SKILL_PATH]. 2. Boundary markers: The instructions do not define delimiters or ignore markers for the content being processed. 3. Capability inventory: The agent is tasked with invoking the skill-reviewer agent and writing modifications (fixes) to the targeted skill files. 4. Sanitization: No sanitization, validation, or filtering of the ingested content is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 07:10 AM
Security Audit — agent-trust-hub — skill-improver