custom-skills-plan-analyze
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from user-provided files (plans, RFCs, reports) which serves as a potential surface for indirect prompt injection.
- Ingestion points: The skill reads the target document provided via the
@pathargument and identifies/reads additional code files mentioned within that document to establish context. - Boundary markers: The instructions do not specify the use of clear delimiters or instructions for the agent to treat the document content strictly as data, which increases the risk of the agent obeying instructions embedded within the analyzed text.
- Capability inventory: The agent has the capability to read local files (source code and documents) and write back to those files in Stage 6 (appending analysis results).
- Sanitization: There is no mention of filtering, escaping, or sanitizing the content of the external files before the agent processes them for analysis or performs file-write operations.
Audit Metadata