The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or safety bypass instructions were detected. The skill focuses on improving agent task management through organized file-based persistence.
[COMMAND_EXECUTION]: The skill includes two helper scripts (init-session.sh and check-complete.sh) for initializing project files and verifying task status. These scripts perform standard filesystem tasks using grep and cat with no network activity or sensitive data access.
[PROMPT_INJECTION]: The skill uses a pattern where research findings are written to files that the agent later reads. This creates an indirect prompt injection surface.
Ingestion points: task_plan.md, findings.md, and progress.md are read during execution to refresh context.
Boundary markers: Content is organized under markdown headers; no specific delimiters are used to isolate instructions.
Capability inventory: Writing to the local filesystem and executing the included shell scripts.
Sanitization: Not present; the agent is expected to summarize external research into these files.

planning-with-files