positioning-mapper
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including user-provided interview notes, win-loss reasons, and competitor messaging scraped from the web. This creates an attack surface where malicious instructions embedded in external content could potentially influence the agent's behavior or output. The instructions include a specific security warning to treat such inputs as untrusted, which serves as a procedural mitigation.
- [COMMAND_EXECUTION]: The skill interacts with local system components by calling
registry-events.pyto submit proposed updates to the claims ledger (claims.ndjson) and launch registry (launches.ndjson). This involves executing local script logic to handle data discovered during the positioning mapping process. - [DATA_EXFILTRATION]: The skill utilizes network-enabled scripts, specifically
firecrawl.py(for web scraping) andtavily.py(for web searching), to retrieve competitor information. While these tools are used for their intended purpose of market research, they represent a capability for network communication and external data retrieval.
Audit Metadata