The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several local Node.js scripts (configure-keyapi-auth.mjs, search-keyapi-docs.mjs, keyapi-api.mjs) to automate workflows. These scripts use standard Node.js built-in modules to interact with the file system and network.
[DATA_EXFILTRATION]: The skill communicates with the vendor's API at api.keyapi.ai to perform Instagram data lookups and content analysis. This behavior is documented and inherent to the skill's primary purpose.
[EXTERNAL_DOWNLOADS]: The skill fetches live documentation, endpoint summaries, and OpenAPI specifications from docs.keyapi.ai to ensure requests match current API contracts.
[PERSISTENCE_MECHANISMS]: The authentication script modifies shell configuration files (such as .bashrc, .zshrc, or PowerShell profiles) to persist the KEYAPI_TOKEN environment variable. This allows the credential to be reused across different shell sessions. The script uses a managed block and robust quoting to handle these modifications safely.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it retrieves and processes untrusted content from Instagram (e.g., post captions and comments).
Ingestion point: scripts/keyapi-api.mjs retrieves external content from the Instagram platform.
Boundary markers: The skill instructions explicitly command the agent to separate observed API facts from inference and to use analytical language.
Capability inventory: The skill can execute local Node.js scripts and has the ability to write API responses to local files via the --output-file argument.
Sanitization: The helper scripts perform standard JSON parsing and stringification, but the instructions rely on the model's analytical guidelines rather than hardcoded content filtering.

keyapi-instagram