incident-postmortem-drafter
Pass
Audited by Gen Agent Trust Hub on Jul 3, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external JSON data and interpolates it into a Markdown report template. The
_md_escapefunction inpostmortem_drafter.pyonly performs whitespace normalization and does not escape Markdown syntax (such as headers, links, or markdown-formatted instructions). This creates a surface for indirect prompt injection if the resulting document is subsequently processed by an AI agent or automated system. - Ingestion points: The
postmortem_drafter.pyscript reads incident metadata and timeline data from an external JSON file specified by the--inputflag. - Boundary markers: The script does not utilize delimiters or specific boundary markers to isolate untrusted input data from the Markdown template structure.
- Capability inventory: The skill possesses file-system read and write capabilities, used to load the input JSON and save the generated Markdown report.
- Sanitization: Sanitization is limited to basic whitespace stripping and line-ending normalization; it lacks validation or escaping for Markdown-specific control characters.
Audit Metadata