The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill functions by reading and writing to an external issue tracker (such as GitHub Issues or a local markdown tracker). This creates a vulnerability to indirect prompt injection if an attacker places malicious instructions in issue titles, bodies, or comments.
Ingestion points: The skill loads the 'map body' and individual 'ticket' bodies/comments from the repository's issue tracker (referenced in SKILL.md under 'The map body' and 'Work through the map').
Boundary markers: The instructions lack explicit boundary markers or instructions for the agent to ignore potentially malicious content within retrieved issue data.
Capability inventory: The skill can create, assign, and close issues, and it can invoke subagents for research and prototyping tasks based on ticket types.
Sanitization: There is no evidence of sanitization or filtering of the external data before it is interpolated into the agent's decision-making flow.
[COMMAND_EXECUTION]: The skill instructions advise the agent to prompt the user to run an external setup command, /setup-matt-pocock-skills, if the required issue tracker environment is not found. This directs the user to execute external configuration code that is not defined or audited within this skill package.

wayfinder