The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute local Python scripts (notion_api.py, notion_browser.py) and the playwright-cli utility. These commands are essential for the skill's workflow, which involves querying the Notion API and automating browser sessions for users who lack API access.
[DYNAMIC_EXECUTION]: The skill implements dynamic code execution by generating JavaScript snippets in notion_browser.py and executing them via playwright-cli eval. This mechanism is used to probe the Notion Document Object Model (DOM) and perform synthetic paste operations to insert content into the editor without requiring clipboard permissions.
[INDIRECT_PROMPT_INJECTION]: The skill reads data from external sources (Notion pages) during its verification and probing steps. While this is necessary for its function, it introduces a vulnerability surface where malicious content within a Notion page could attempt to influence the agent's context or behavior.
Ingestion points: Data enters the agent's context through API responses in notion_api.py (cmd_probe, cmd_readback) and browser DOM text extraction in notion_browser.py (probe-js, verify-js).
Boundary markers: The skill requires a mandatory dry-run (Step 3) where the user reviews the draft before writing, providing a manual verification layer. However, it lacks explicit delimiters to isolate the data read back from Notion.
Capability inventory: The agent has access to Bash (shell execution), urllib.request (network operations via Python), and playwright-cli (browser automation).
Sanitization: Content inserted via browser automation is sanitized using json.dumps before being interpolated into a JavaScript string literal, preventing basic script injection into the automation logic.

notion-report