skill-improver
Pass
Audited by Gen Agent Trust Hub on May 17, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes external improvement plans to perform file modifications.
- Ingestion points: Improvement plan markdown files (improvement-plan-{name}.md) are read into context in SKILL.md.
- Boundary markers: No delimiters or specific instructions are used to prevent the agent from obeying malicious directions embedded within the 'Suggested content' or 'Reason' fields of the plan.
- Capability inventory: The skill utilizes Edit and Write tools to modify any file specified in the plan and executes shell scripts (SKILL.md).
- Sanitization: Content validation is absent for the data being written to files.
- [COMMAND_EXECUTION]: The skill invokes local bash scripts to perform backups and integrity checks.
- Evidence: SKILL.md specifies the execution of scripts/backup-skill.sh and scripts/verify-update.sh using shell commands.
Audit Metadata