dfg-inc/praxis-ba/accept
Claude Code and agent skill by dfg-inc/praxis-ba
- community
Checking that the SKILL.md file is still in the repository.
Use once a plan-approved WP's development harness reports its validation is green — the sole human-gated path to baselining a WP's requirements, the only verb in this system that ever writes baselines/.
Install
npx vskill@latest install dfg-inc/praxis-ba/acceptSKILL.md
Accept
accept is the single baseline writer — "there is no standalone baseline verb, only accept writes baselines." It is transactional, not authoring: the WP and its FR/NFR/BR closure already exist and are batched/plan-approved; this verb flips them to baselined, freezes their content into baselines/BL-…/, and resolves any CR whose full declared impacts set (amends and/or spawns entries) is now provably delivered — every amends entry's target carries a ## History entry citing the CR, every spawns entry's realized target's traces_to includes the CR.
Preconditions
- WP is
status: plan-approved(prepare-wp/approve-planalready ran). - The development harness's validation is green (whatever that harness defines — this skill only gates on the evidence JSON below; it never runs, or even needs to know, what that harness's own checks were).
- A VERIFY-EVIDENCE JSON exists (
{wpId, commit, testRunHashes, suites, producedAt, producedBy, toolVersion}, typically.ba/cache/verify-<wp-id>.json) whosecommitmatches the currentHEADand whosetestRunHashesare present — the honest, non-forgery-proof signalacceptGatechecks; it does not re-run anything itself.
Human gate
Accepting = baselining = the point of no easy return (a baselined item's content can only change again via shape-requirement's Case B, a full CR-gated version bump). This is an explicit human-only action (CLAUDE.md: "Human-only actions: … accept …").
- Summarize to the human what is about to be baselined: the WP, its FR/NFR/BR closure, the evidence file, and whether every
acceptGatecheck (vision confirmed, AC-presence, plan-approved, evidence matches HEAD) is currently green. - AskUserQuestion: "Accept and baseline
<wp-id>now?" — never infer a yes from a green evidence file alone; the human decision is the gate, the evidence file is only input to it. - On yes:
`` praxis-ba accept --repo <canon-dir> --wp <id> --evidence <path-to-verify-evidence.json> --date <YYYYMMDD> --head-commit <sha> ` (--date/--head-commit are optional — omit to let the CLI default to today / git rev-parse HEAD under --repo.) Writes VERIFY-OK/VERIFY-FAIL, never a bare PASS — a FAIL here means the gate rejected the accept, not that anything was partially baselined (the transaction is crash-safe and idempotent via a write-ahead marker; a legitimate crash mid-transaction self-resumes on the next accept` call for the same WP).
Done when
accept returns VERIFY-OK, the WP is status: accepted, its delivered set is baselined under a new baselines/BL-…/, and a human explicitly triggered it — never a machine-initiated baseline.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingplugins/ba/skills/accept/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/dfg-inc/praxis-ba/accept)