portal-app-init
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to read, interpret, and merge content from user-supplied target repositories into an AGENTS.md file, which constitutes an indirect prompt injection surface.
- Ingestion points: Reads the existing
AGENTS.mdand configuration files (manifest, README) from the target repository supplied by the user. - Boundary markers: Utilizes
formalcode fences and specific ODC DSL syntax (RULE,MUST,MUST_NOT) to delineate instructions. - Capability inventory: The skill is capable of modifying files (
AGENTS.md) and references instructions for using browser-based testing tools (Playwright, Computer Use, Chrome extensions). - Sanitization: Implements semantic merging and re-verification steps (
@verificationrule) to ensure new conventions are integrated without duplicating or overwriting project-specific facts. - [SAFE]: All external references, such as
@xgent/*packages and related skills (xgent-app-release,portal-dev-setup), are associated with the internalxgent-aivendor ecosystem. The skill explicitly forbids the inclusion or hardcoding of credentials and focuses on documenting best practices for local development and performance.
Audit Metadata