implementing-features
Execute ONE approved plan artifact — code changes strictly within the plan's scope, verified by the plan's Verification commands — then stop at a…
Skill by hierovision
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/hierovision-implementing-features
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: documents, design, writing, productivity
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Node.js package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/implementing-features
What's inside
- Skill name
implementing-features- Description
- Execute ONE approved plan artifact — code changes strictly within the plan's scope, verified by the plan's Verification commands — then stop at a manual-validation handoff. Use whenever the user says "implement the plan", "build feat-x", "execute the plan at .opencode/plans/<slug>.md", "the plan is approved, go", or hands off from a design pass — even without saying "implement" or "build". Refuses scope creep (cites the plan's Excluded list; records follow-ups). Mechanical plan-vs-reality mismatches proceed with a History note; contract-breaking ones stop and route back to designing-architecture. Not for trivial single-file changes with no plan (just edit), authoring or revising plans (that is designing-architecture), or deep debugging (separate skill).
- License
- MIT
- Instructions
- 251 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Agent Skills spec issues
- files: referenced but missing: scripts/watch_agent.py
Files (58)
- SKILL.md
- evals.json · evals/
- feat-offline-queue-broken.md · evals/fixtures/phasewave/.opencode/plans/
- feat-offline-queue-rename.md · evals/fixtures/phasewave/.opencode/plans/
- feat-offline-queue.md · evals/fixtures/phasewave/.opencode/plans/
- AGENTS.md · evals/fixtures/phasewave/
- ROADMAP.md · evals/fixtures/phasewave/
- schema.sql · evals/fixtures/phasewave/db/
- timer.spec.ts · evals/fixtures/phasewave/e2e/
- package.json · evals/fixtures/phasewave/
- lint.js · evals/fixtures/phasewave/scripts/
- test.js · evals/fixtures/phasewave/scripts/
- type-check.js · evals/fixtures/phasewave/scripts/
- .gitkeep · evals/fixtures/phasewave/src/lib/
- focus-timer.ts · evals/fixtures/phasewave/src/stores/
- client.ts · evals/fixtures/phasewave/src/supabase/
- feat-clip-duration.md · evals/fixtures/phasewave/tdd-realgap/.opencode/plans/
- AGENTS.md · evals/fixtures/phasewave/tdd-realgap/
- package.json · evals/fixtures/phasewave/tdd-realgap/
- lint.js · evals/fixtures/phasewave/tdd-realgap/scripts/
- test.js · evals/fixtures/phasewave/tdd-realgap/scripts/
- type-check.js · evals/fixtures/phasewave/tdd-realgap/scripts/
- .gitkeep · evals/fixtures/phasewave/tdd-realgap/src/lib/
- clip.js · evals/fixtures/phasewave/tdd-realgap/src/stores/
- feat-audit-log-dedupe.md · evals/fixtures/phasewave/tdd-rebalance/.opencode/plans/
- AGENTS.md · evals/fixtures/phasewave/tdd-rebalance/
- package.json · evals/fixtures/phasewave/tdd-rebalance/
- lint.js · evals/fixtures/phasewave/tdd-rebalance/scripts/
- test.js · evals/fixtures/phasewave/tdd-rebalance/scripts/
- type-check.js · evals/fixtures/phasewave/tdd-rebalance/scripts/
- seed.js · evals/fixtures/phasewave/tdd-rebalance/
- audit-store.broken.js · evals/fixtures/phasewave/tdd-rebalance/src/stores/
- audit-store.js · evals/fixtures/phasewave/tdd-rebalance/src/stores/
- fake-client.js · evals/fixtures/phasewave/tdd-rebalance/src/supabase/
- feat-session-score.md · evals/fixtures/phasewave/tdd-redfirst/.opencode/plans/
- AGENTS.md · evals/fixtures/phasewave/tdd-redfirst/
- package.json · evals/fixtures/phasewave/tdd-redfirst/
- lint.js · evals/fixtures/phasewave/tdd-redfirst/scripts/
- test.js · evals/fixtures/phasewave/tdd-redfirst/scripts/
- type-check.js · evals/fixtures/phasewave/tdd-redfirst/scripts/
- .gitkeep · evals/fixtures/phasewave/tdd-redfirst/src/lib/
- focus-timer.js · evals/fixtures/phasewave/tdd-redfirst/src/stores/
- feat-session-score.md · evals/fixtures/phasewave/tdd-wrongreason/.opencode/plans/
- AGENTS.md · evals/fixtures/phasewave/tdd-wrongreason/
- package.json · evals/fixtures/phasewave/tdd-wrongreason/
- lint.js · evals/fixtures/phasewave/tdd-wrongreason/scripts/
- test.js · evals/fixtures/phasewave/tdd-wrongreason/scripts/
- type-check.js · evals/fixtures/phasewave/tdd-wrongreason/scripts/
- .gitkeep · evals/fixtures/phasewave/tdd-wrongreason/src/lib/
- focus-timer.js · evals/fixtures/phasewave/tdd-wrongreason/src/stores/
- make-session.js · evals/fixtures/phasewave/tdd-wrongreason/tests/helpers/
- database.types.ts · evals/fixtures/phasewave/types/
- coverage-gate.md · references/
- deviation-protocol.md · references/
- handoff-template.md · references/
- red-first.md · references/
- skills-library.md · references/stacks/
- vue-supabase.md · references/stacks/
SKILL.md (first lines)
# Implementing Features
Execute one approved plan artifact. The plan is a contract: implement what
`Files to Modify` + `Included` scope say, after a red-first test per
Acceptance Criterion (Step 5), a runtime validation of the changed flow
(Step 8), and a coverage-and-quality gate (Step 9); run the plan's
`Verification` until exit 0, append a `## History` entry, then **stop** at a
manual-validation handoff. Success is exit codes, never intent — and a
criterion that needs human eyes is never self-certified.
## The implement pass
Copy this checklist and check off items as you complete them.
```
Implement Progress:
- [ ] 1. Read the plan artifact (frontmatter status first)
- [ ] 2. Resolve scope contract: Files to Modify + Included + Excluded
- [ ] 3. Detect stack & load the matching stack reference
- [ ] 4. Reconcile plan vs the actual repo (deviation triage BEFORE edits)
- [ ] 5. RED — capture core behavior from Acceptance Criteria (one test per AC, before any edit; confirm red for the right reason)
- [ ] 6. Implement exactly Files to Modify (scope discipline)
- [ ] 7. Run the plan's Verification; fix + re-run until all exit 0
- [ ] 8. Runtime UI validation (invoke validating-ui; console net + UX review; bounded fix loop)
- [ ] 9. Coverage-and-quality gate (rebalance layer guesses; re-prove meaningfulness; expand only on a real gap)
- [ ] 10. Update the plan (History entry; status per the user's convention)
- [ ] 11. Completion handoff summary + MANUAL validation steps; STOP
```
### Step 1 — Read the plan artifact
The skill starts only when given a plan artifact path. Read it before touching
any source — the prompt alone is not the spec; the plan is.
Git setup first (per `reference/git-workflow.md`): confirm the working tree is
on the plan's branch — `<type>/<name>` derived from the slug (e.g.
`fix-activity-image-delete` → `fix/activity-image-delete`), created from a
current `main` if it does not exist. Never work on `main`; only PRs merge.
- **No plan path given**, work is multi-file or schema-touching → route to
`designing-architecture` and stop. This skill does not reverse-engineer a
plan from a prompt; skipping the design stage destroys the auditability.
- **No plan path given**, single-file trivial change → just edit it; a plan
costs more than the change (see "When not to use").
Read the frontmatter `status:` next:
- `approved` → proceed.
- `draft` → not approved. **Stop and confirm with the user** before editing
("the plan at `<path>` is `status: draft` — treat as approved, or hand back
to design?"). Do not silently treat a draft as approved.
- `superseded` → do not implement; follow the `related:` pointer to the
superseding plan.
A plan missing `Files to Modify`, `Verification`, or `Scope` is a contract
defect — route back to `designing-architecture` rather than improvising.
### Step 2 — Resolve the scope contract
The scope contract is three lists, read together:
1. **Files to Modify** — the only files you may create or edit; a file not on
the list is out of scope. Do not tidy, refactor, or fix a pre-existing bug
in it on the way past.
2. **Scope → Included** — what the plan delivers; confirms intent when a
per-file note is ambiguous.
3. **Scope → Excluded** — the boundary you cite when refusing scope creep
(non-empty by plan-format contract).
Hold the three in working memory for the whole pass; every "should I do this
too?" is checked against them.
### Step 3 — Detect stack & load the matching stack reference
Identify the project stack from its rules file (`AGENTS.md` /
`.opencode/agents.md`). If a matching `references/stacks/<stack>.md` exists in
this skill's own directory (not the project's), read it now and apply its
implementation-time concerns. If none exists, proceed generically and flag the
gap in the handoff. The available stack references are listed under
**References**.
### Step 4 — Reconcile plan vs the actual repo (BEFORE editing)
Before editing, match every `Files to Modify` entry against what is on disk and
read each named file. Classify each mismatch immediately:
- **Mechanical** — the feature is unchanged; only a mechanical detail differs
(file moved/renamed, import alias drift). Proceed at the corrected detail and
record a dated `## History` note with the reason.
- **Contract-breaking** — the mismatch changes what is being built (a different
feature, a schema differing from `Schema / Type Impacts`, an AC untestable as
written). **STOP**. Do not edit source. Report the mismatch and route back to
`designing-architecture`; the implementer never improvises a different design.
The dividing question is *does the difference change the feature being built?*
When in doubt, stop — a stop is recoverable, a built-wrong-feature is not. The
full classification table is in `references/deviation-protocol.md` (read it
when a mismatch is ambiguous).
### Step 5 — RED — capture core behavior from Acceptance Criteria
Before any source edit, author each testable AC as a failing test — one test
per AC, before the behaviour exists. Invoke the matching test-trio skill in its
**red-first, pre-implementation** mode (the right-layer table lives there).
**Run it: it must fail for the right reason** — the failure output names the
missing behaviour, not a fixture gap or a typo; fix the test's own setup until
it does. Record each AC → test file and the red evidence for the plan History
and the handoff at Step 11.
Honest stops: an AC whose verifier cannot produce a red-first test is
contract-breaking (Step 4); a red that never fails for the right reason is the
N-attempts stop (Step 7) — hand to `debugging-test-failures`. Do not proceed on
an unverified red; that silently turns red-first back into test-after.
Full contract, rationale, and the pre- vs post-implementation distinction:
`references/red-first.md`.
### Step 6 — Implement exactly Files to Modify (scope discipline)From hierovision/ai-framework, MIT.
More from hierovision/ai-framework
- 0.592score
auditing-supply-chain
hierovision
Audit a software supply chain — dependency manifests and lockfiles, build provenance, SBOMs, and release artifacts — for what makes it untrustworthy…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
refining-issue-acceptance
hierovision
Refine a GitHub issue into a proper engineering spec — a short, clear description plus well-defined acceptance criteria (AC) written as observable, external…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
reviewing-code
hierovision
Review a change (a diff, a branch, or an implement pass's output) against the plan that authorized it and against engineering quality, then issue a verdict…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
debugging-test-failures
hierovision
Diagnose WHY a failing verification won't converge — reproduce first, hypothesis + discriminating experiment, classify the defect (code / test / plan /…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
capturing-ui-evidence
hierovision
Captures the current visual state of UI targets into a structured, addressable EVIDENCE ARTIFACT — screenshots (full-page + element-clipped) plus a…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
deploying-to-azure-swa
hierovision
Deploy a static front end to Azure Static Web Apps from GitHub Actions — preview-per-PR environments, production promotion on push to main, capturing the…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
deploying-with-supabase
hierovision
Deploy to Supabase from CI — `supabase db push` forward-only semantics, "only one db push at a time" concurrency, migrations in CI via session-mode `--db-url`…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
designing-architecture
hierovision
Turn ONE selected roadmap entry or direct feature request into a verifiable implementation plan that a separate build session can execute cold — then stop at…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
designing-cicd
hierovision
Design GitHub Actions CI/CD topology — trigger selection (push / pull_request / workflow_run / workflow_dispatch), job sequencing via needs, concurrency…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
facilitating-product-vision
hierovision
Facilitates an owner-led product vision session and produces two artifacts: a strategy-only vision document plus roadmap items (outcome themes under…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
handling-security-incidents
hierovision
Handle a security incident through the NIST SP 800-61 lifecycle — validate and scope the signal, build a timestamped fact timeline, classify severity, plan…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
managing-github-issues
hierovision
Persist an established ROADMAP.md as GitHub issues — one issue per open roadmap row, enriched from its source files, tagged with a stable roadmap-id label for…
SkillAFreshness: A★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 7
- Last commit (this folder)
- Oct 16, 2026 (13 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/hierovision/ai-framework
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from hierovision
- 0.592score
auditing-supply-chain
hierovision
Audit a software supply chain — dependency manifests and lockfiles, build provenance, SBOMs, and release artifacts — for what makes it untrustworthy…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
designing-cicd
hierovision
Design GitHub Actions CI/CD topology — trigger selection (push / pull_request / workflow_run / workflow_dispatch), job sequencing via needs, concurrency…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
deploying-to-azure-swa
hierovision
Deploy a static front end to Azure Static Web Apps from GitHub Actions — preview-per-PR environments, production promotion on push to main, capturing the…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
deploying-with-supabase
hierovision
Deploy to Supabase from CI — `supabase db push` forward-only semantics, "only one db push at a time" concurrency, migrations in CI via session-mode `--db-url`…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
designing-architecture
hierovision
Turn ONE selected roadmap entry or direct feature request into a verifiable implementation plan that a separate build session can execute cold — then stop at…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
facilitating-product-vision
hierovision
Facilitates an owner-led product vision session and produces two artifacts: a strategy-only vision document plus roadmap items (outcome themes under…
SkillAFreshness: A★ 1Claude Code +9
Cite this page
implementing-features skill. Agent Mart. https://agentmart.dev/skills/hierovision-implementing-features (last updated 2026-10-24)