continuous-learning

Pass

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • Indirect Prompt Injection (LOW): The skill is designed to ingest data from user development sessions to generate instructions and clustering into new skills.
  • Ingestion points: Development session context and coding patterns extracted via 'PreCompact' and 'Stop' hooks as mentioned in SKILL.md.
  • Boundary markers: Absent; no mention of delimiters or instructions to ignore embedded code-level injections during pattern extraction.
  • Capability inventory: File-write operations to .claude/instincts/ and logic for skill generation/evolution.
  • Sanitization: Absent; the documentation does not describe any validation or escaping of extracted patterns before they are used to 'evolve' into skills.
  • Data Exposure (LOW): The skill automates the extraction and local storage of development patterns to .claude/instincts/. While no network exfiltration is hardcoded, the /instinct-export command provides a mechanism for sharing potentially proprietary or sensitive architectural decisions.
  • No Code (SAFE): The provided skill consists solely of markdown documentation and YAML frontmatter. No executable scripts, binaries, or configuration files were included for analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 20, 2026, 08:25 PM