The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes file system operations including reading project documentation (Read), searching for files (Glob), and creating directories and files (mkdir, Write). It also interacts with the Linear API (save_project, update_project, get_project) to manage project states. These operations are performed to facilitate the core functionality of coordinating project Epics and are consistent with the skill's stated purpose.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection during its Phase 1 research step and Phase 4 preparation step. It automatically scans and reads content from markdown files (e.g., docs/requirements.md) and all .html files in the repository to extract business goals and functional domains. This automated processing of potentially untrusted data creates a pathway for hidden instructions in the documents to influence the agent's behavior during the decomposition process. Ingestion points: Local markdown project documents and frontend HTML source files. Boundary markers: Absent; the skill does not use delimiters to isolate ingested data from agent instructions. Capability inventory: Linear project management API access and local file system write permissions. Sanitization: No validation or sanitization logic is implemented for the data extracted from external files.

ln-210-epic-coordinator