show-me-your-work
Pass
Audited by Gen Agent Trust Hub on Jul 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a bash script (scripts/log.sh) that uses mkdir and printf to manage a local TSV log file. This is a standard administrative task for the skill's purpose.
- [PROMPT_INJECTION]: The skill instructions direct the agent to read agent-transcripts/ to verify its decision log. This constitutes an indirect prompt injection surface. 1. Ingestion points: agent-transcripts/ directory (SKILL.md). 2. Boundary markers: None specified for the transcript reading process. 3. Capability inventory: Local file writing via scripts/log.sh. 4. Sanitization: scripts/log.sh implements input sanitization to prevent spreadsheet formula injection, though it does not specifically filter for prompt injection patterns.
- [SAFE]: The skill demonstrates defensive programming by sanitizing inputs in scripts/log.sh to prevent spreadsheet formula injection. It also explicitly warns the agent against accessing unrelated private transcripts in other project directories (~/.cursor/projects/*/).
Audit Metadata