The Agent Skills Directory

[SAFE]: The skill facilitates project tracking through local markdown files and standard git operations. The provided script scripts/resume.sh performs read-only operations using grep and git log to summarize project status, presenting no risk of unauthorized execution or data exfiltration. All referenced templates and documentation focus solely on developer workflow management.
[INDIRECT_PROMPT_INJECTION]: The skill processes project-level planning documents (SESSION.md, IMPLEMENTATION_PHASES.md) to maintain state. While this involves processing external data, the activity is the primary purpose of the skill and is managed within the user's local repository.
Ingestion points: The skill reads SESSION.md and docs/IMPLEMENTATION_PHASES.md from the project root directory.
Boundary markers: Content is organized using structured markdown headers, though no explicit adversarial instruction delimiters are employed.
Capability inventory: The skill's operations are limited to standard git commits and file read/write actions within the project workspace.
Sanitization: The skill relies on natural language parsing of the session files without explicit input sanitization filters.

project-session-management