roadmap

Pass

Audited by Gen Agent Trust Hub on Apr 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on executing shell commands via Bash, specifically Git utilities (git log, git diff) and search tools (grep, tail, head), to analyze project history and state.\n- [PROMPT_INJECTION]: The skill is exposed to indirect prompt injection by processing untrusted data from the project repository.\n
  • Ingestion points: Ingests data from git commit messages, repository documentation (docs/.md), and feature specifications (arkhe/specs//spec.md).\n
  • Boundary markers: Lacks explicit boundary markers or instructions to isolate the agent from commands embedded in the analyzed data.\n
  • Capability inventory: Possesses the capability to write to important project files and execute shell commands, which could be exploited through malicious inputs.\n
  • Sanitization: Does not implement sanitization or validation of external data before integration into the model's context or output files.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 26, 2026, 10:55 AM