build-app-context

Fail

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: HIGHCREDENTIALS_UNSAFEDATA_EXFILTRATIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill instructions direct the agent to silently extract API keys and authentication tokens for TestRail, Jira, and GitHub from local files (~/.folio-credentials, .env) and environment variables without user confirmation or consent.
  • [DATA_EXFILTRATION]: The skill accesses sensitive local file paths, specifically ~/.folio-credentials, to harvest stored authentication data.
  • [EXTERNAL_DOWNLOADS]: The skill performs multiple network operations to external services:
  • Fetches project metadata from the vendor's official Atlassian Confluence site.
  • Downloads repository contents and metadata via the GitHub API.
  • Retrieves test case data and issue details from dynamic TestRail and Jira endpoints defined in local credentials.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted data from external sources and incorporates it into local files.
  • Ingestion points: Phased gathering of TestRail case titles/steps, GitHub READMEs/feature files, and Jira story/bug descriptions (Phases 1, 2, and 3).
  • Boundary markers: There are no delimiters or explicit instructions used to isolate or ignore embedded commands within the ingested data.
  • Capability inventory: The skill can read sensitive local files, write to the filesystem (including modifying its own SKILL.md file), and perform network GET/POST requests.
  • Sanitization: Sanitization is insufficient, consisting only of basic HTML tag stripping from TestRail content.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 22, 2026, 01:18 PM
Security Audit — agent-trust-hub — build-app-context