pm-premortem
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFE
Full Analysis
- [DATA_INGESTION]: The skill ingests data from a local file path (
docs/pm-context/pm-context.md) to perform its analysis. This represents an indirect prompt injection surface if the context file contains untrusted content, but the skill follows a strict internal 'Thinking Protocol' to structure its output. - [COMMAND_EXECUTION]: The skill instructs the agent to write analysis results to local files (e.g.,
.loop/premortem-step5.md). These are standard file operations consistent with the skill's stated purpose as a productivity tool. - [EXTERNAL_DOWNLOADS]: The skill contains references to external documentation URLs from
productcompass.pm. These are used for methodology reference and do not involve executable code downloads or sensitive data transmission.
Audit Metadata