harness-planner

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run bash agent/init.sh to resolve dependency closures and establish the initial state.
  • [PROMPT_INJECTION]: The skill identifies and ingests data from multiple files—including agent/vision.md, agent/harness/horizon.json, and agent/categories.json—to influence its decision-making and planning process.
  • Ingestion points: The agent reads state, vision, and category data from local filesystem paths to build and score candidate sets.
  • Boundary markers: The instructions lack explicit delimiters or instructions for the agent to ignore potentially malicious embedded content within these ingested files.
  • Capability inventory: The skill has the capability to execute shell scripts and write plan artifacts to the filesystem.
  • Sanitization: There is no evidence of sanitization or validation logic for the content read from external files before it is processed by the LLM.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 07:16 PM
Security Audit — agent-trust-hub — harness-planner