continuous-learning

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill defines a continuous learning workflow that ingests data from external sources (coding sessions, diffs, and user feedback), creating a surface for indirect prompt injection.
  • Ingestion points: The agent is instructed to review session changes, diffs, and user corrections recorded in the session context (SKILL.md).
  • Boundary markers: Absent; no specific delimiters or instruction-ignore warnings are defined to separate ingested code content from the agent's logical instructions.
  • Capability inventory: The agent is granted permission to perform file-write operations to update local project documentation, specifically targeting files like CLAUDE.md, LEARNED.md, and the knowledge/ directory (SKILL.md).
  • Sanitization: Absent; the skill does not define methods for escaping or validating ingested patterns and feedback before they are incorporated into the long-term knowledge base.
  • [NO_CODE]: The skill consists exclusively of markdown-based instructions and process definitions. It does not include any executable scripts, binaries, or remote package dependencies.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 12:53 PM
Security Audit — agent-trust-hub — continuous-learning