ask-project-memory

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill reads from .docs/decisions.md to guide agent behavior and ensure consistency with past decisions. Maliciously crafted instructions in this file could potentially influence the agent's reasoning, though the skill lacks high-risk capabilities like network access or arbitrary command execution.
  • Ingestion points: .docs/decisions.md (referenced in README.md and SKILL.md).
  • Boundary markers: None present.
  • Capability inventory: Local file read and write operations restricted to documentation paths.
  • Sanitization: No explicit sanitization of the markdown content is performed before the agent processes it.
  • [Data Exposure] (SAFE): The skill is designed to interact with project documentation. It does not access sensitive system paths, credentials, or perform network exfiltration.
  • [Command Execution] (SAFE): While the skill includes a Python script (scripts/validate.py), it is a benign placeholder and does not execute untrusted commands or download remote resources.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 10:03 AM