skills/affaan-m/ecc/strategic-compact/Gen Agent Trust Hub

strategic-compact

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill relies on analyzing the session transcript to calculate token usage, creating a potential surface for indirect prompt injection as the transcript includes untrusted external data and user-provided content.\n
  • Ingestion points: The skill reads usage records from the session transcript provided via the transcript_path in the hook payload.\n
  • Boundary markers: The instructions do not specify any delimiters or protective prompts to differentiate between control data and user content within the transcript.\n
  • Capability inventory: The skill executes shell commands via Node.js as part of the PreToolUse hook mechanism.\n
  • Sanitization: The documentation lacks mention of input validation or sanitization for the data retrieved from the transcript.\n- [COMMAND_EXECUTION]: The skill is designed to run a local Node.js script (suggest-compact.js) using shell hooks triggered during agent activities (Edit and Write). While this is a standard integration pattern for the platform, it grants the skill the ability to execute code automatically based on agent operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 12:06 PM
Security Audit — agent-trust-hub — strategic-compact