learnings
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted output from multiple sources to distill and store learnings. Ingestion points include output from
git status,git diff,pytest, and other shell command execution logs. No boundary markers or delimiters are specified to separate untrusted data from the agent's internal instructions. The skill has the capability to write to the file system and execute local binaries, but it provides no evidence of sanitization or validation of the ingested content. - [EXTERNAL_DOWNLOADS]: The skill installs the
learningsdependency from the author's Homebrew tap. Evidence: Therun_learnings_toolfunction inSKILL.mdperformsbrew install tkersey/tap/learningswhen the binary is not detected. - [COMMAND_EXECUTION]: Spawns system processes to gather environment context and manage the learnings database. Evidence: Executes
git status -sb,git diff --stat, and local binariesappend_learningandlearningswith user-influenced arguments.
Audit Metadata