sdlc-write-stories

Pass

Audited by Gen Agent Trust Hub on Nov 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and process untrusted data from the project wiki to generate user stories and commit them back to the repository.
  • Ingestion points: The skill reads external content from overview.md, prd.md, and change-prd.md files within the target wiki repository to identify functional requirements, actors, and goals.
  • Boundary markers: The instructions do not specify the use of delimiters or explicit system instructions to isolate the ingested requirement text from the agent's control logic, which could allow maliciously crafted requirements to influence agent behavior.
  • Capability inventory: The agent has the capability to write new Markdown files, modify repository logs, and execute git commands (branch, commit, and push) based on the processed data.
  • Sanitization: There is no evidence of sanitization, filtering, or escaping of the input data before it is interpolated into the generated user story files.
  • Mitigation: The risk is significantly reduced by the inclusion of three mandatory approval gates (Intent, Design, and Change-set) where a human must review the proposed changes before the agent proceeds.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 13, 2026, 09:20 AM