The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from project-specific files to guide agent behavior, creating a vulnerability surface where malicious content in those files could influence the agent's actions.
Ingestion points: The agent is instructed to read IMPLEMENTATION_PHASES.md during project initialization and SESSION.md during session resumption to determine the current state and next actions (SKILL.md).
Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to ignore potentially malicious instructions embedded within the progress trackers or planning documents.
Capability inventory: The skill possesses capabilities to write files, execute git commits, and run a local reporting script (resume.sh).
Sanitization: There is no evidence of content sanitization or validation of the 'Next Action' field or task descriptions before they are used to direct the agent's focus.
[COMMAND_EXECUTION]: The skill involves the execution of local shell commands and scripts to manage session state.
Script Execution: The skill includes scripts/resume.sh, which uses grep, sed, and git to extract and display status information from SESSION.md.
Git Operations: The agent is instructed to perform git commits (git commit) using a structured checkpoint format which includes state information derived from the project environment (SKILL.md, templates/checkpoint-commit-format.md).

project-session-management