project-skill-protocol

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest user-defined rules and store them as "overlays" that are automatically injected into the context of other skills during execution.
  • Ingestion points: User-supplied rule text provided during the add and update operations as described in SKILL.md Phase 2 and 3.
  • Boundary markers: The skill implements a mandatory "PROPOSAL GATE" (SKILL.md Phase 2, step 9) requiring explicit user approval before writing rules to disk. It also uses "additive-only" phrasing to distinguish project rules from core framework protocols.
  • Capability inventory: The skill possesses file-write capabilities to the docs/project-protocols/ and docs/project-reference/ directories, and it reads other skills' protocol files to check for contradictions.
  • Sanitization: The skill enforces an "additive-only screen" (references/registry.md §4) that explicitly refuses any rules instructing the agent to skip, replace, or ignore framework protocols or safety gates. It also subjects all rules to a prompt-engineering rubric and the $prompt-enhance tool to ensure instructions are observable and imperative.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 08:16 PM