skills/autumnsgrove/lattice/druid/Gen Agent Trust Hub

druid

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill implements a workflow that involves reading existing instruction files from the .claude/skills/ directory and using that content to design and write new skills, creating a multi-step instruction chain.
  • Ingestion points: In the 'COMMUNE' phase (SKILL.md), the skill reads frontmatter and core purposes from every SKILL.md file located in the .claude/skills/ subdirectories to build an ecosystem map.
  • Capability inventory: The skill instructions (SKILL.md) direct the agent to create directories and write new markdown instruction files to the .claude/skills/ and docs/scratch/ directories using shell commands like mkdir -p and file write operations.
  • Boundary markers: The ritual phases lack explicit boundary markers or instructions to treat the ingested data as untrusted, which could lead the agent to follow instructions embedded within the analyzed skills during the creation of new ones.
  • Sanitization: The skill does not define validation or sanitization rules for the naming of new creatures ({concept}, {name}) or for the content derived from existing skills, which is a necessary step to prevent path manipulation or the propagation of malicious instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 01:31 AM