github-archive

Pass

Audited by Gen Agent Trust Hub on Jun 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Recommends the installation of standard, trusted Google Cloud libraries (google-cloud-bigquery and google-auth) from official package registries to facilitate BigQuery interactions.
  • [DATA_EXPOSURE]: Provides guidance on utilizing a service account JSON file for authentication. While it references a sensitive file path (path/to/credentials.json), it is used as a standard configuration placeholder for legitimate Google Cloud SDK usage.
  • [COMMAND_EXECUTION]: Provides Python code snippets and SQL templates to execute queries against the public githubarchive dataset. The skill includes commendable security best practices, such as cost estimation dry-runs and maximum_bytes_billed limits to prevent unintentional resource exhaustion.
  • [DATA_EXFILTRATION]: The skill facilitates the retrieval of GitHub event data from BigQuery. While this involves data movement from the archive to the agent's context, it is the primary purpose of the skill and targets well-known Google Cloud infrastructure.
  • [INDIRECT_PROMPT_INJECTION]: The skill inherently processes untrusted data (user-generated content from GitHub issues, PRs, and commit messages stored in the archive).
  • Ingestion points: Data retrieved from BigQuery payload fields in SKILL.md investigation patterns.
  • Boundary markers: None explicitly provided in the templates to delimit retrieved content from agent instructions.
  • Capability inventory: The skill utilizes the BigQuery client to fetch and process external data.
  • Sanitization: No specific sanitization or filtering logic is provided for the retrieved text content, representing a standard vulnerability surface for forensic analysis tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 7, 2026, 01:29 PM