The Agent Skills Directory

[Command Execution] (SAFE): The skill utilizes Bash to execute standard Git commands such as git status, git log, and git branch. These operations are well-scoped to the primary purpose of identifying project state.
[Indirect Prompt Injection] (LOW): The skill reads external project documentation and task trackers which could be used as a vector for instructions to influence the agent.
Ingestion points: Reads files from .claude/blueprints/prds/, docs/blueprint/feature-tracker.json, and .claude/blueprints/work-orders/.
Boundary markers: Absent. No instructions are provided to the agent to disregard embedded directives within these files.
Capability inventory: Read, Bash, Grep, Glob, Edit, Write.
Sanitization: Absent. Data is processed directly as project context.

project-continue