The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill lacks sanitization for data ingested from external sources, making it vulnerable to indirect prompt injection.
Ingestion points: The skill reads from a local TASKS.md file and extracts tasks from meeting summaries or conversation history.
Boundary markers: Absent. There are no instructions to ignore embedded commands or use specific delimiters for data vs. instructions.
Capability inventory: The skill has file system read/write capabilities within the current working directory.
Sanitization: Absent. The skill summarizes and processes content directly without filtering.
[COMMAND_EXECUTION]: The skill instructs the agent to perform file operations and references a specific tool command.
Evidence: The skill directs the agent to copy dashboard.html from ${CLAUDE_PLUGIN_ROOT}/skills/ to the current working directory.
Evidence: The instructions mention a /productivity:start command for system setup.

task-management