SF Civic Digest

Track San Francisco city government activity — Board of Supervisors, Land Use committee, Public Safety committee, SFMTA Engineering Public Hearings, SF Plann...

This skill appears to do what it says: scrape public SF government and journalism sources and produce JSON for an agent to turn into a digest. Before you run it: (1) inspect the scripts locally and run them in a sandboxed environment; they perform many network requests and will write archive/state files into the skill directory; (2) ensure runtime dependencies are present (python3, and if you need PDF parsing the environment should have pdftotext or Python pdf libraries like pdfplumber); (3) be prepared for long runs (the master script may make ~20+ network calls and take minutes); (4) confirm you are comfortable with automated scraping per site terms of service and rate limits; (5) if you plan to let the agent invoke this autonomously, consider limiting frequency or running it in an environment with restricted network/file privileges. If you want higher confidence, provide the omitted source files for review (the manifest showed many truncated files) and confirm whether any scripts require non-standard Python packages or external binaries.

Type: OpenClaw Skill Name: sf-civic-digest Version: 1.0.0 The sf-civic-digest skill is a comprehensive data aggregation tool for monitoring San Francisco government activity. It uses standard Python libraries (urllib, subprocess, re) to fetch and parse public data from official sources like Socrata, Legistar, and various city department websites. The scripts are well-documented, follow a consistent architecture, and include state management for tracking updates. There is no evidence of data exfiltration, malicious execution, or prompt injection; the 'editorial lens' described in STYLE.md is a persona configuration for the agent rather than a security threat. The code even includes basic sanitization in scripts like sf_311.py to prevent SoQL injection.