narrative-tracker
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
geo-tracker.pyusing a shell command to persist narrative snapshots and scores. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external AI responses and interpolates it directly into a shell command context without sanitization or boundary markers, creating a vulnerability surface for command injection. \n
- Ingestion points: AI-generated responses captured from platforms like ChatGPT, Perplexity, Gemini, and others (SKILL.md, Step 2). \n
- Boundary markers: Absent; there are no delimiters or instructions provided to the agent to ignore or escape embedded commands within the processed data. \n
- Capability inventory: The skill has the capability to execute shell commands via the
pythoninterpreter (SKILL.md, Step 5). \n - Sanitization: Absent; the
{what the AI said}variable is used directly in a shell command string without escaping or validation.
Audit Metadata