The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions and strategy files direct the agent to execute shell commands such as curl, jq, grep, and agent-browser to perform network requests and parse data from external sources.
[EXTERNAL_DOWNLOADS]: The skill fetches data from multiple external platforms and APIs, including Bluesky, Mastodon, Reddit, and third-party scraping services like ScrapeCreators and Apify.
[DATA_EXFILTRATION]: The skill performs network operations to non-whitelisted domains (such as api.scrapecreators.com and api.apify.com) to process user-supplied social media URLs and API keys.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from various social media platforms (X, LinkedIn, Instagram, etc.), which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Social media post content, author metadata, and reply threads fetched via external URLs across multiple platforms.
Boundary markers: The provided documentation does not define specific delimiters or instructions for the agent to treat fetched social media content as untrusted data.
Capability inventory: The skill possesses network access via curl, browser control via agent-browser, and the ability to write fetched data to the local file system in ~/Documents/social-fetches/.
Sanitization: The skill normalizes external data into a structured JSON schema, which provides a layer of data structure but does not necessarily sanitize natural language instructions within the text fields.

social-fetch