algolia-crawler
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from user-specified web URLs. Ingestion points: External web URLs (SKILL.md, workflow.md). Boundary markers: None explicitly defined in the extraction templates to delimit content from instructions. Capability inventory: Execution of algolia crawler CLI commands to write to an index. Sanitization: Content is cleaned using Cheerio/JS string methods but not specifically checked for embedded prompt instructions. This surface is inherent to the skill's purpose.
- [DYNAMIC_EXECUTION]: The skill involves creating a JavaScript recordExtractor function which is stringified and included in a JSON configuration file. Evidence: references/workflow.md shows the use of jq to embed the source code of extractor.js into a config.json file.
- [CREDENTIALS_UNSAFE]: The skill handles sensitive credentials, including Crawler API keys and Algolia Write keys. The skill follows safe practices by advising users to manage these via environment variables or local configuration files. Evidence: references/cli.md provides instructions for exporting credentials as environment variables.
- [EXTERNAL_DOWNLOADS]: The skill references the installation of the official Algolia CLI via Homebrew. Evidence: references/cli.md points to the official algolia/algolia-cli repository.
Audit Metadata