skills/affaan-m/ecc/ecc-council/Gen Agent Trust Hub

ecc-council

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a set of prompting instructions for decision synthesis. It does not contain executable scripts or binary files.
  • [SAFE]: The workflow involves spawning isolated subagents with limited context, which acts as a security boundary against context-based anchoring.
  • [SAFE]: The persistence section explicitly instructs the agent not to write to sensitive or shadow directories like ~/.claude/notes, instead recommending standard project management tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests codebase-specific snippets and issue text, representing a potential surface for instructions embedded in data. However, as the skill only outputs text-based reasoning and does not perform file-system writes or network actions based on this input, the risk is minimal.
  • Ingestion points: Codebase files, snippets, and issue text referenced in Step 2.
  • Boundary markers: The prompt template for subagents in Step 4 uses clear structural separators ("Question:", "Context:", "Respond with:") to define role boundaries.
  • Capability inventory: The skill only launches sub-prompts for reasoning; it lacks capabilities for subprocess execution, network requests, or direct file-system modification.
  • Sanitization: No explicit sanitization or escaping of ingested context is performed within the prompt template.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 02:52 PM