The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill explicitly instructs the agent to gather "Test credentials if relevant" and "Credentials: [if needed for testing]" to be included in the continuation notes file.
[DATA_EXFILTRATION]: The skill workflow includes committing and pushing the continuation file (which the skill suggests should contain credentials) to a remote repository via the git push command.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with the local git repository (git add, git commit, git push) and the system clipboard (pbcopy, xclip) to move data between session contexts.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted state information from the current project environment and interpolates it into a resume prompt for a future agent session.
Ingestion points: The skill reads session state, file modification history, and git commit logs from the project directory (SKILL.md).
Boundary markers: No delimiters or safety instructions are provided to distinguish between the gathered state data and the structure of the resume prompt.
Capability inventory: The skill possesses the ability to write markdown files, execute git commands, and write directly to the system clipboard.
Sanitization: The skill does not validate or sanitize the gathered project information before including it in files or the clipboard buffer.

handoff