Skill Claude CodeCodex
What this project is allowed to claim in public — the anchoring rule for AI-generated summaries, confidence states, attribution, and the wording of every published number. Use this skill whenever writing an LLM prompt, rendering generated prose, labelling a metric, designing a badge or status, drafting social posts…
Skill Claude CodeCodex
Hard infrastructure ceilings for this project — GitHub API rates, GitHub Actions behaviour, Cloudflare Pages build quotas, Groq limits, and forbidden platforms. Use this skill before choosing any hosting provider, database, scheduler, or LLM provider; before changing polling frequency or the number of watched…
Skill Claude CodeCodex
Skill "instrument-ui" from sighttrue/sighttrue, covering the thesis, banned patterns, the near-miss to avoid, density is the flex and typography.
Skill Claude CodeCodex
How the data agent fetches, stores, and schedules GitHub signals for this project — the tiered 4-hourly cadence, conditional requests, live state versus historical ledger, spike baselines, and failure tolerance. Use this skill whenever writing or modifying any collector, the GitHub Actions workflow, the ledger schema…
Instructions file
Claude Code instructions for sighttrue/sighttrue, covering sighttrue — unified developer signal agent, non-negotiables, skill routing by phase, cadence and storage layout.
Skill Claude CodeCodex
Take a dated reading before adopting a dependency, base image, runtime version, AI model or hosting provider. Use when adding or upgrading a package, choosing a Docker base image or language version, picking a model, or answering whether something is still maintained, still supported, or still shipping.