prompt-improver
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-supplied prompts and agent instruction documents (such as plan files or system instructions) to generate improved versions. This creates an attack surface where malicious instructions embedded within the input data could influence the agent's behavior during the analysis phase or be propagated into the resulting 'Improved Prompt'.
- Ingestion points: The skill takes a user-provided prompt string as primary input and explicitly mentions reading local documentation files like 'plan files' or 'AGENTS.md' as improvement targets in SKILL.md.
- Boundary markers: The instructions do not specify the use of clear delimiters (like XML tags or triple quotes) or provide warnings for the agent to ignore instructions embedded within the content being analyzed.
- Capability inventory: The skill is intended for use in environments like Claude Code and Codex CLI, which possess significant capabilities including file system access (read/write), shell execution, and the ability to run development tools (e.g., pnpm, rg, apply_patch).
- Sanitization: There is no mention of sanitizing, escaping, or filtering the input content before it is processed by the agent.
Audit Metadata