The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes bash commands to manage files and check tool availability. Evidence includes commands such as which agcron, ls .cron/jobs/, cat .cron/jobs/{id}.md, rm .cron/jobs/{id}.md, and cat > .cron/jobs/{id}.md for file operations. Explicit user confirmation is required for both write and delete operations.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of reading and writing external files that define agent jobs. Ingestion points: The skill reads content from .cron/jobs/{id}.md using the cat command in SKILL.md. Boundary markers: There are no explicit markers used to separate external file content from the agent's internal instructions when displayed. Capability inventory: The skill has the ability to write to files using redirected output (cat >) and delete files using rm. Sanitization: Input validation is performed for fields like cron expressions and adapter names as detailed in SKILL.md and references/field-reference.md.

editing-jobs