skills/erphq/skills/docs-site/Gen Agent Trust Hub

docs-site

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill describes a standard operational playbook for building and deploying documentation. It interacts with the official API of build.host, which is a resource owned by the vendor (erphq).
  • [CREDENTIALS_UNSAFE]: The skill references the sensitive credential file located at ~/.build-host/credentials.json. It mitigates risk by providing strict instructions that the agent must never print API keys, OAuth tokens, or secrets in source code, logs, or chat messages.
  • [COMMAND_EXECUTION]: The operating procedure involves running local build checks and project commands. This is a standard requirement for site generation tools and is limited to the user's project directory.
  • [PROMPT_INJECTION]: The skill processes user project data and source code to generate documentation, which is a surface for indirect prompt injection.
  • Ingestion points: Reads files from the local project directory and git repository (SKILL.md).
  • Boundary markers: None explicitly defined to isolate input data from instructions.
  • Capability inventory: Includes file system writes, local command execution for build checks, and network requests to the build.host API.
  • Sanitization: The skill does not document specific sanitization or filtering logic for the processed project data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 04:43 PM
Security Audit — agent-trust-hub — docs-site