The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources, creating a surface for indirect prompt injection.
Ingestion points: The skill fetches content from external URLs, including Atom/RSS feeds, GitHub repositories, and general web pages as described in the Subscription Types section of SKILL.md.
Boundary markers: There are no explicit instructions or delimiters provided to the agent to ignore potentially malicious instructions embedded within the fetched feed content or web pages.
Capability inventory: The skill uses the joelclaw CLI for resource management and relies on an Inngest-based backend (packages/system-bus/src/inngest/functions/subscriptions.ts) to trigger LLM-based summarization (subscription/summarize).
Sanitization: The skill instructions do not specify any sanitization, filtering, or validation steps for the external content before it is passed to the LLM for summarization.
[COMMAND_EXECUTION]: The skill uses a specialized CLI tool, joelclaw, to perform its primary functions. These commands are executed within the agent's environment to list, add, remove, and check subscriptions.
[EXTERNAL_DOWNLOADS]: The skill is intended to interact with external web resources, including GitHub APIs and various feed formats (RSS/Atom), to monitor for changes.

monitor