@dzhechkov/skills-meta

The canonical meta-skill pack — the development-process skills, packaged as Agent Skills for distribution and cross-platform use.

The pack contains 21 development-process skills:

| Skill | Files | Purpose | |---|---|---| | feature-adr | 34 | ADR-driven feature development pipeline (per-stage model routing incl. Codex, cross-model QE, usage-adaptive pre-emptive Codex switch at ≥70% Claude usage; amendment-Confirmation discipline, derived 🚦 Gates line, opt-in Step-10 Delivery Gate, Step-8 no-stubs gate — an unwaived TODO-class marker in the run's touched files is a HIGH gap; waivers require a reason) | | design-tokens | 1 | DESIGN.md design-token contract — lint (WCAG AA) / diff / export Tailwind+DTCG (wraps @google/design.md) | | knowledge-extractor | 10 | Harvest reusable knowledge from a project (/harvest) | | audit | 4 | Whole-codebase 5-phase deep audit-and-act (recon → self-challenge → P0-P3 → approval → fix) | | goap-research-ed25519 | 6 | GOAP-planned research with ed25519-signed provenance | | design-thinking | 5 | Stanford d.school 5-phase + validate product discovery | | agentshield-scan | 4 | Security scanner for .claude/ agent configurations | | adversarial-verifier | 4 | Adversarial claim verification / paranoid fact-checking | | capture-adr | 4 | Lightweight mid-conversation ADR capture (MADR 4.0) | | context-window-management | 4 | Priority-based pruning + checkpoint/restore under context pressure | | external-comms-gate | 4 | Screens outbound comms for confidential-info leaks | | reflection-loop | 4 | Standalone critique → revise self-review cycle | | session-recap | 1 | Mid-session status protocol: mandatory refresh, then exactly four points (the requester's words, proof rather than green tests, human-vs-machine blockers, and owner-tagged next steps), with no new work during the recap. Distinct from dz recap, the records-only calendar retrospective. | | skill-advisor | 4 | Recommends which skills / presets / packages fit a task (meta-advisor) | | skill-crystallizer | 4 | Auto-creates / combines / repairs skills from traces | | structured-reasoning | 4 | Reasoning-strategy router (ToT / CoT / compression) | | understand-anything-bridge | 4 | Bridge to the Understand-Anything knowledge-graph tooling | | explore | 3 | Socratic task clarification | | problem-solver-enhanced | 1 | First-principles + TRIZ + game-theory problem solving | | loop-plan-author | 2 | Author agent LOOPS as typed loop-plan/1 plans — four patterns, INV-1..8, claims/defers, USER regions, reading a run's trace. Packaged for Claude Code as @dzhechkov/loop-designer-plugin; dz AUTHORS, GATES and READS loops — it never RUNS one | | decision-mockups | 10 | Owner-facing DECISION PAGE — plain-language explanation, browser-frame before/after mockups, clickable option forks and a copy-answers export that pastes back into the session, plus a zero-dependency deterministic gate (references/check_page.py, G0–G14) that refuses a page with a fake fork, an untokenised colour, an external resource or a leftover placeholder. Vendored mirror — the canon (and the test suite) lives in @dzhechkov/skills-decision-mockups; heal with dz sync-canonical decision-mockups --from packages/@dzhechkov/skills-decision-mockups/decision-mockups. Honest scope: the skill writes Russian (its gate hard-requires the export literals) |

goap-research-ed25519 — self-learning (optional, since this release). When @dzhechkov/harness-cli is on PATH, the bundled research skill recalls prior METHOD lessons at the start of an investigation and records new ones at four named moments. Without it the skill behaves exactly as before and says so once — it is detected, never required. Lessons go to a SEPARATE store (<project>/.health-brain/.dz) and never to the shared one; recall reads both, so engineering lessons transfer in and medical ones do not leave. A format check refuses identifier shapes (email, phone, record numbers) — it does NOT judge whether a lesson describes a method or a person, and says so: that judgement is the agent's, per the teach protocol. See skills/goap-research-ed25519/SKILL.md.

Install

# Via dz CLI (recommended) — the current meta preset; session-recap wiring is deferred
dz init --target claude-code --preset meta

# Or pick specific skills
dz init --target claude-code --select audit,skill-advisor,feature-adr

# Or install the package directly
npm install @dzhechkov/skills-meta

Canonical vs legacy — the coexistence model

Each canonical skill here is projected into the repository's .claude/skills/<name>/ tree by the root sync command. This is deliberate (ADR-002); session-recap follows the same projection path:

• .claude/skills/ — the legacy tree Claude Code reads today. Preserved, untouched (ADR-001).
• packages/@dzhechkov/skills-meta/ — the canonical source, ready to be published as an npm package and compiled to any platform.

During the coexistence period the two are kept in agreement by scripts/sync-canonical-to-legacy.mjs, which compiles each canonical skill with @dzhechkov/adapter-claude and compares the result to the legacy tree. While the copies match, sync is a no-op — it never modifies a legacy skill.

Status

next — carries the canonical feature-adr gate and module with requirement coverage (C8), ADR ids by in-file heading, NOT-ESTABLISHED on an unclosed fence in a declaration input, the planner prompt naming its inputs, and ONE bracketed plan-repair round (backup → snapshot → repair → snapshot → re-gate → archive or proven restore). Staged: not yet versioned or published. See @dzhechkov/skills-feature-adr for the full account.

0.9.51 — carries the canonical feature-adr module with the new mandatory Observability section in Step 5: how the shipped feature will be watched, with "nothing to observe" accepted as a complete answer. See @dzhechkov/[email protected].

0.9.50 — carries the canonical feature-adr skill with the C6 amendment-integrity fix (each amendment is scoped to its own block instead of a three-line window, closing both a false refusal and a false pass) and the Step-6 planner instruction that the amendment row is machine-read. See @dzhechkov/[email protected] for the full account.

0.9.44 — a fix to the vendored feature-adr gate script, no new skills.

The bundled K2 plan-completeness gate now tells an ABSENT 00_complexity_assessment.md from a deliberate skip: a missing file is a WARNING that names the missing artifact, while a present file carrying no acid table is an honest SKIP. Both used to report as a clean skip, so the acid check silently switched itself off whenever Step 0 wrote nothing — MEASURED 2026-08-21, 66 of 199 features carried that file and the last four in a row did not.

0.9.42 — adds decision-mockups (19 → 20 skills): the owner-facing decision page — plain language, options with a stated price and a named recommendation, and a copy-answers button whose output is meant to be pasted straight back into a chat. Vendored from @dzhechkov/skills-decision-mockups (the canonical home; heal this mirror with dz sync-canonical decision-mockups --from packages/@dzhechkov/skills-decision-mockups/decision-mockups — a bare sync resolves THIS pack as canonical and would overwrite the real home). Lands together with @dzhechkov/[email protected], which adds the same id to the meta preset.

0.9.5 — part of the extended-a-migration feature. Additive only: these are copies; the originals in .claude/skills/ are never moved or modified.

How to use

Skills auto-activate — your agent loads a skill when your task matches its trigger phrases (defined in each skill's SKILL.md frontmatter). For example:

• "Audit the codebase" (or /audit) → audit
• "Which skill should I use for X?" → skill-advisor
• "Implement feature X" → feature-adr

To see a skill's exact triggers and assets: dz info <skill-id>. To find one: dz registry search <term>.

Write discipline — why the writing steps now build a skeleton first

MEASURED in the field on 2026-08-19 and 2026-08-20, two independent runs of the canonical pipeline against a ~130-file repository: the steps that must produce a document — Step 5 (Architecture) and Step 6 (Plan) — never reached a write. 18 attempts, zero file writes in every one. They died in the READING phase: a shell result returns, the agent thinks about its next move, stays silent past the runtime's 180-second inactivity watchdog, and is killed. One run cost ~4M tokens and 1h54m.

The failure is deterministic, not unlucky: thinking time grows with accumulated history, so on a large enough repository unbounded exploration guarantees the kill. And the cause was in the instructions, not the runtime — every writing step said, in effect, "read the code, write the document", with no reading budget and no order of operations. An agent obeying literally reads until it dies.

The owner's control experiment is the whole evidence base, and it is n=1: same slice, same inputs, same model, ONE added paragraph about write discipline → a 10-section skeleton on disk 8 minutes in, first attempt, after six consecutive deaths.

So every document-producing step now carries:

1. the FIRST file write happens within the first ~12 tool calls — a skeleton of section headings with one line of intent under each;
2. then fill it one section per edit, no edit longer than ~120 lines;
3. never go more than 2 minutes without a tool call;
4. when unsure whether to read more or to write — write.

What this does not do, stated plainly: it does not stop the deaths. The watchdog is unchanged. It changes what survives one — previously nothing, now a skeleton on disk that the next attempt can continue. The ~12 and ~120 are chosen parameters, not measured optima. Two possible regressions were named by an independent reviewer and are not measured: an early skeleton may anchor a structure chosen before understanding, and section-per-edit raises tool-call overhead.

Steps whose deliverable is a returned verdict rather than a document (the complexity router) are deliberately excluded — "skeleton first" is nonsense there.

0.9.45 — the bundled feature-adr Step-8 module now describes the amendment gate as the command dz amendment-check instead of a judgement the reviewer is asked to make. Text only; no pipeline change.

Signature scope (this release)

The pack's .dz-manifest.json now covers exactly the files this package SHIPS, as reported by npm pack — not everything present in the author's working tree. Previously it signed files that files[] excludes (typically CHANGELOG.md), so every recipient's verifier reported listed in the manifest but absent and the pack read as TAMPERED. Re-signing at any earlier moment could not fix that: those files were never in the tarball.

Nothing about the shipped content changed in this release — only what the signature describes.

Shared Markdown masking in feature-adr gates

The standalone plan-completeness gate ships with markdown-masker.mjs, copied byte-for-byte from harness-core's src/markdown-masker.ts. It runs without a core build. Amendment checks, swarm briefs and K2 share the parser while retaining their existing unclosed-block and indentation policies. The four-space indented-code gap remains open for amendment checks and K2; swarm briefs retain their existing masking of indented code. Versions are unchanged in this staged change.