feat: add CI workflow, CONTRIBUTING.md, CHANGELOG.md, and README badges/demo#1
Conversation
…es/demo Agent-Logs-Url: https://github.com/Smithbox-ai/ControlFlow/sessions/4a8567c4-a4c2-41c3-91d1-be277236ebbd Co-authored-by: Smithbox-ai <192372536+Smithbox-ai@users.noreply.github.com>
Agent-Logs-Url: https://github.com/Smithbox-ai/ControlFlow/sessions/4a8567c4-a4c2-41c3-91d1-be277236ebbd Co-authored-by: Smithbox-ai <192372536+Smithbox-ai@users.noreply.github.com>
There was a problem hiding this comment.
Pull request overview
Adds project “community traction” essentials (CI + contributor/release docs) and improves README first-impression by surfacing status badges and a quick-start workflow explanation.
Changes:
- Add GitHub Actions CI workflow to run
cd evals && npm teston pushes/PRs. - Update README with badges and a new “How It Works” section emphasizing the 3-step workflow.
- Add new
CONTRIBUTING.mdandCHANGELOG.mdwith guidance and a v1.0.0 entry.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.
| File | Description |
|---|---|
.github/workflows/ci.yml |
Introduces CI job that installs eval deps and runs eval test suite. |
README.md |
Adds badges and a concise “How It Works” onboarding section. |
CONTRIBUTING.md |
Documents contribution workflows (running evals, adding/editing agents, skills, PR expectations). |
CHANGELOG.md |
Adds Keep-a-Changelog formatted v1.0.0 entry capturing system capabilities. |
| **Eval suite (302 checks)** | ||
|
|
||
| - Pass 1: Schema validity (Ajv strict mode, JSON Schema 2020-12) | ||
| - Pass 2–3: Scenario integrity and cross-scenario structural regression (179 structural checks) | ||
| - Pass 4: P.A.R.T section order enforcement | ||
| - Pass 4b: Clarification trigger and tool routing section validation | ||
| - Pass 5: Skill library registration integrity | ||
| - Pass 6: Synthetic rename negative-path checks | ||
| - Pass 7: Prompt behavior contract behavioral regression (74 checks across 9 agents) | ||
| - Pass 8: Orchestration handoff contract regression (49 checks) |
There was a problem hiding this comment.
The changelog claims the eval suite has “302 checks” with a specific breakdown, but evals/README.md currently documents a different total. Consider either aligning these numbers with the authoritative source or avoiding fixed counts in the changelog to prevent staleness.
| **Eval suite (302 checks)** | |
| - Pass 1: Schema validity (Ajv strict mode, JSON Schema 2020-12) | |
| - Pass 2–3: Scenario integrity and cross-scenario structural regression (179 structural checks) | |
| - Pass 4: P.A.R.T section order enforcement | |
| - Pass 4b: Clarification trigger and tool routing section validation | |
| - Pass 5: Skill library registration integrity | |
| - Pass 6: Synthetic rename negative-path checks | |
| - Pass 7: Prompt behavior contract behavioral regression (74 checks across 9 agents) | |
| - Pass 8: Orchestration handoff contract regression (49 checks) | |
| **Eval suite** | |
| - Pass 1: Schema validity (Ajv strict mode, JSON Schema 2020-12) | |
| - Pass 2–3: Scenario integrity and cross-scenario structural regression | |
| - Pass 4: P.A.R.T section order enforcement | |
| - Pass 4b: Clarification trigger and tool routing section validation | |
| - Pass 5: Skill library registration integrity | |
| - Pass 6: Synthetic rename negative-path checks | |
| - Pass 7: Prompt behavior contract behavioral regression across agent prompts | |
| - Pass 8: Orchestration handoff contract regression |
| branches: ["master"] | ||
| pull_request: | ||
| branches: ["master"] |
There was a problem hiding this comment.
The workflow is only triggered for the master branch. The repo/diff context suggests the default branch may be main, in which case CI won’t run for pushes/PRs to the default branch and the README CI badge may show no results. Consider triggering on both main and master, or updating the branch filter to match the repo’s actual default branch.
| branches: ["master"] | |
| pull_request: | |
| branches: ["master"] | |
| branches: ["main", "master"] | |
| pull_request: | |
| branches: ["main", "master"] |
|
|
||
| [](https://github.com/Smithbox-ai/ControlFlow/actions/workflows/ci.yml) | ||
|  | ||
|  |
There was a problem hiding this comment.
The “Eval Checks” badge hardcodes 302, but the repository’s evals/README.md currently states a different total (283). Hardcoding the number risks the badge and docs drifting out of sync; consider removing the fixed count or deriving it from a single authoritative source and updating all references together.
|  | |
|  |
| npm test | ||
| ``` | ||
|
|
||
| All 302 checks must pass before any PR can be merged. The suite runs fully offline. |
There was a problem hiding this comment.
This section states “All 302 checks must pass…”, but evals/README.md currently documents a different total (283). To avoid documentation drift, consider removing the hardcoded number (e.g., “All eval checks must pass”) or updating both files to match the authoritative count.
| All 302 checks must pass before any PR can be merged. The suite runs fully offline. | |
| All eval checks must pass before any PR can be merged. The suite runs fully offline. |
|
|
||
| ```bash | ||
| cd evals | ||
| npm install |
There was a problem hiding this comment.
The contributor instructions use npm install, while CI uses npm ci. Using npm ci locally (with the committed lockfile) better matches CI’s deterministic dependency resolution and reduces “works locally but not in CI” issues.
| npm install | |
| npm ci |
Repository lacked CI, contributor guidance, release history, and a compelling first impression — all friction points that reduce community traction.
Changes
.github/workflows/ci.ymlcd evals && npm teston every push/PR tomasterpermissions: contents: read(least-privilegeGITHUB_TOKEN)README.md@Planner→ approve →@Orchestrator) instead of architecture prose — reduces time-to-value for new visitorsCONTRIBUTING.mdCHANGELOG.md