Skills

A skill is a package of structured files that teaches an AI coding agent how to work with a specific tool or framework. The skill below was generated by Great Docs from this project’s documentation. Install it in your agent and it will be able to run commands, edit configuration, write content, and troubleshoot problems without step-by-step guidance from you.

Any agent — install with npx:

npx skills add https://yasyf.github.io/cc-steer/

Codex / OpenCode

Tell the agent:
Fetch the skill file at https://yasyf.github.io/cc-steer/skill.md and follow the instructions.

Manual — download the skill file:

curl -O https://yasyf.github.io/cc-steer/skill.md

Or browse the SKILL.md file.

SKILL.md

---
name: cc-steer
description: >
  Your best training data is rotting in ~/.claude. Use when writing Python code that uses the cc_steer package.
license: PolyForm-Noncommercial-1.0.0
compatibility: Requires Python >=3.13.
---

# cc-steer

Your best training data is rotting in ~/.claude.

## Installation

```bash
pip install cc-steer
```

## API overview

### Scanning

Mine feedback events out of transcripts, incrementally.

- `scan.scan`
- `scan.ScanReport`

### Detection

The signal detectors and the candidate shape they emit.

- `detectors.detect`
- `models.FeedbackCandidate`
- `models.dedup_key`

### Storage

The feedback event store.

- `store.FeedbackStore`

### Retraining

The weekly retrain lanes — gate and watcher — and their promotion machinery.

- `retrain.lexical.retrain_gate`
- `retrain.watcher.retrain_watcher`
- `retrain.watcher.WatcherRecipe`
- `retrain.watcher.register_watcher_adapter`
- `retrain.watcher.seed_incumbent_probs`
- `retrain.evalset.freeze_eval`
- `retrain.evalset.freeze_steer_type`
- `retrain.evalset.freeze_pick`
- `retrain.evalset.EvalFrame`
- `retrain.evalset.SteerTypeFrame`
- `retrain.evalset.PickFrame`
- `retrain.promotion.corrected_gate`
- `retrain.promotion.GateResult`
- `retrain.promotion.Verdict`

### Many-shot watcher

Score the frozen frame through a byte-constant, cache-amortized many-shot exemplar prefix.

- `retrain.manyshot.build_exemplar_system`
- `retrain.manyshot.score_frame`

### Encoder gate

Fine-tune a transformer encoder as the stage-1 gate and score it paired on the frozen gate frame.

- `retrain.encoder.EncoderSpec`
- `retrain.encoder.EncoderModel`
- `retrain.encoder.train_encoder`
- `retrain.encoder.score_frozen`

### Instrument

Paired fast-DeLong AUC comparison and the instrument-card decision rule.

- `instrument.delong_se`
- `instrument.paired_delong`
- `instrument.PairedDeLong`
- `instrument.bootstrap_ci`
- `instrument.mde`
- `instrument.actionable`
- `instrument.InstrumentCard`
- `instrument.Comparison`
- `instrument.paired_verdict`
- `instrument.unpaired_verdict`
- `retrain.sweep.ArmScore`
- `retrain.sweep.compare_score_reports`

### Decisions

Mine AskUserQuestion decision rounds from transcript trees into a parquet dataset.

- `decisions.mine`
- `decisions.DecisionRow`
- `decisions.MineResult`
- `decisions.Quarantine`
- `decisions.write_decisions`
- `decisions.read_decisions`
- `decisions.stats_of`
- `decisions.DecisionStats`

## Resources

- [Full documentation](https://yasyf.github.io/cc-steer/)
- [llms.txt](llms.txt) — Indexed API reference for LLMs
- [llms-full.txt](llms-full.txt) — Comprehensive documentation for LLMs
- [Source code](https://github.com/yasyf/cc-steer)