capture-learning
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by ingesting and persisting untrusted session content to the filesystem.
- Ingestion points: Session Q&A context used in Step 3 to generate the markdown document.
- Boundary markers: The skill does not implement delimiters or specific warnings to ignore embedded instructions within the captured content.
- Capability inventory: The skill has the capability to write files to both the project-local directory and the user's home directory archive.
- Sanitization: No sanitization or filtering is applied to the ingested session content before it is written to the filesystem.
- [COMMAND_EXECUTION]: The skill executes local system commands to retrieve metadata about the project environment.
- Evidence: Step 2 instructs the agent to read the 'git remote' of the current project to determine the repository name.
Audit Metadata