The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The README instructions recommend installation using piped shell scripts fetched from the author's GitHub repository. While this is a high-risk pattern, the scripts are hosted on the author's own GitHub repository and are required for the tool's installation.
Evidence: curl -fsSL https://raw.githubusercontent.com/dozybot001/Intent/main/scripts/install.sh | bash
Evidence: irm https://raw.githubusercontent.com/dozybot001/Intent/main/scripts/install.ps1 | iex
[COMMAND_EXECUTION]: The CLI tool executes local Git commands to manage history state and retrieve repository information.
Evidence: Usage of subprocess.run for git rev-parse, git status, and git remote in src/intent_cli/store.py.
[EXTERNAL_DOWNLOADS]: The skill downloads its core logic from GitHub and uses standard tools like pipx for installation.
[DATA_EXFILTRATION]: The hub sync command transmits semantic history data to a user-configured remote API. This is the primary method for sharing history across different environments.
Evidence: http_json POST requests in src/intent_cli/commands/hub.py.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes data from the local repository and semantic history files.
Ingestion points: Reads JSON history objects from the .intent/ directory structure (src/intent_cli/store.py).
Boundary markers: Instructions explicitly direct the agent to parse structured JSON output rather than interpreting raw text from the CLI tool.
Capability inventory: The agent can execute shell commands and perform network data synchronization.
Sanitization: Data is handled through structured JSON serialization.

intent-cli