The Agent Skills Directory

[COMMAND_EXECUTION]: The skill leverages the github-mcp-repo-writes-enabled tool to push new files and updates to the Noisemaker111/skills repository on GitHub. This functionality is the primary intended use case for the skill owner to manage their bot's capabilities.- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for generating new instructions based on untrusted data from the conversation history.
Ingestion points: Recent conversation history ('a flow just done') is used as the template for new skills (SKILL.md).
Boundary markers: The skill does not specify any delimiters or safety instructions to prevent the agent from obeying malicious instructions embedded within the analyzed flow.
Capability inventory: The agent possesses write access to the author's GitHub repository via the github-mcp-repo-writes-enabled tool.
Sanitization: There is no requirement for the agent to sanitize or validate the content of the flow before converting it into a permanent skill. This creates a surface where an attacker could influence a conversation flow that the user then unknowingly saves as a repeatable command.

bot-commands