pre-mortem
Pass
Audited by Gen Agent Trust Hub on Jun 12, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local shell commands (
ls,grep,head,tail,jq) and a specialized CLI tool (ao) to manage project metadata, search for prior implementation learnings, and verify plan integrity. - [COMMAND_EXECUTION]: It executes a local maintenance script (
hooks/finding-compiler.sh) to refresh compiled findings after updating the local findings registry. - [DATA_EXPOSURE]: The skill reads from and writes to the
.agents/directory, which is the standard location for persistent agent state and historical findings in this environment. - [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it ingests and analyzes implementation plans provided by the user or prior sessions. It mitigates this risk by enforcing a 'no-self-grading invariant,' requiring that validation is performed by a context-isolated 'blind sub-agent' to prevent author bias and instruction leakage.
Audit Metadata