pm-manipulation
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes standard shell commands and utilities to perform its tasks.
- Uses
curlto interact with public Polymarket Gamma and CLOB APIs for market data, price history, and user comments. - Uses
mktempandcatto manage temporary files for report generation. - Executes a local
./notifyscript to deliver findings to the operator. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources which could be used to influence the agent's behavior or reporting.
- Ingestion points: Fetches user comments from Polymarket events (
gamma-api.polymarket.com/comments) and retrieves content from regional press, Telegram, X (Twitter), and Reddit viaWebSearchandWebFetch. - Boundary markers: The skill instructions do not specify the use of delimiters or 'ignore' instructions when processing or summarizing this external content.
- Capability inventory: The agent has the ability to read/write to local memory and output directories, perform network requests via
curland search tools, and execute a notification script. - Sanitization: While the skill applies algorithmic filters to detect bot behavior (regex patterns and frequency analysis), it lacks explicit sanitization or escaping for external text before it is interpolated into the final markdown report and notification briefing.
Audit Metadata