typescript-writing-code
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to process, review, and refactor external TypeScript code, creating an inherent surface for indirect prompt injection.\n
- Ingestion points: User-provided or external TypeScript source code referenced in SKILL.md.\n
- Boundary markers: Absent; there are no specific instructions to the agent to treat input code as untrusted data or to use delimiters.\n
- Capability inventory: The agent is instructed to execute shell commands (pnpm run test, pnpm run lint) in references/post-change-protocol.md which could trigger malicious payloads if present in the input code's test suites.\n
- Sanitization: No sanitization or validation of input code is performed before the verification steps are executed.\n- [COMMAND_EXECUTION]: The skill contains explicit instructions to execute local shell commands for code validation (pnpm, tsc, biome). These are standard development operations and align with the skill's stated purpose of verifying code changes.
Audit Metadata