The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The script scripts/generate.py contains logic to search for and read sensitive authentication profile files belonging to other agents on the local system located at ~/.openclaw/agents/*/agent/auth-profiles.json. It also reads from ~/.config/fal/api_key.
[DATA_EXFILTRATION]: The skill reads sensitive local credential files and subsequently sends the recovered API keys to external services (OpenRouter and Fal) during its operation.
[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection (Category 8).
Ingestion points: The script scripts/generate.py (line 307) fetches arbitrary text content from user-provided URLs using the Jina Reader API.
Boundary markers: The external content is directly interpolated into the LLM prompt (line 258) without any delimiters, XML tags, or instructions to ignore embedded commands.
Capability inventory: The agent has the capability to write files to the local filesystem (plan_path.write_text, image.save, caption_path.write_text) and perform authenticated network requests to external APIs.
Sanitization: No sanitization, escaping, or validation is performed on the content fetched from the remote URL before it is processed by the language model.
[EXTERNAL_DOWNLOADS]: The skill downloads arbitrary web content via the r.jina.ai proxy service based on user-supplied URLs.

instagram-slides