PerryLink/dsh-doublecheck

dsh-doublecheck

Turn "it's done" into an evidence-backed verdict: grill the requirements first, prove the test red then green, and let a four-phase gate settle deliverable or rework before the handoff.

The delivery quality gate for DeepSeek Harness: two plugin rows watch the same replayable session log and turn the grill → design → red → green → review → verify discipline loop into commands, tools, and one verdict that lands in a report.

Install · Configuration · FAQ · 简体中文 · English · Español · Português · हिन्दी

dsh plugin --profile <name> add dsh-doublecheck

Community plugin, not affiliated with DeepSeek. Canonical repository: https://github.com/PerryLink/dsh-doublecheck (Gitee is a mirror).

What it is

dsh-doublecheck does not install a prompt; it installs two plugin rows. doublecheck-grill grills the requirements and commits them as a spec; doublecheck-guard holds the line at every edit, test, and handoff boundary. Both read the same durable session log, so after a resume, a fork, or a restart they see exactly the same evidence.

A /gate run folds what already happened in the session into a four-phase checklist — were the requirements confirmed one by one, is the latest test run red or green, does every edit serve a spec dimension, what does the review conclude — and settles one deliverable / rework required verdict, written as a report you can paste into a PR description.

It does not write your code and it does not judge how the code looks. It does one thing: it makes the sentence "it's done" carry evidence.

Three-step setup

1. Install (pick one channel):

   # git channel (latest main; the prepare script builds with production dependencies only)
   dsh plugin --profile web add github:PerryLink/dsh-doublecheck
   # npm channel (published releases)
   dsh plugin --profile web add dsh-doublecheck
   # tarball channel: pnpm pack in this repo, then
   dsh plugin --profile web add ./dsh-doublecheck-<version>.tgz
   

   Both rows (doublecheck-grill and doublecheck-guard) activate with the profile.

2. Minimal config: none is required — the bundle patch already inserts both rows; an id-targeted override has to restate the whole row's config (see "Configuration"). The equivalent profile-patch shape:

   - insert:
       - id: doublecheck-grill
         name: 'dsh-doublecheck/grill'
       - id: doublecheck-guard
         name: 'dsh-doublecheck/guard'
   

3. Verify it mounts:

   dsh --profile web --dump-config | grep -E -A3 'id: doublecheck-(grill|guard)'
   

   > /doublecheck status
   > /gate status
   

Uninstall: dsh plugin --profile web remove dsh-doublecheck (or remove the two rows from the profile patch).

Want a zero-configuration strict mode — every gate on at block intensity, with coverage evidence required — apply the shipped overlay on top of the bundle patch: dsh --profile web --patch ./node_modules/dsh-doublecheck/strict.patch.yml.

Core capabilities

Every checkpoint in dsh-doublecheck takes its evidence from the session log, not from a claim that "I tested it":

• Requirements furnace (doublecheck-grill) — the bundled grill-requirements skill, the model-facing doublecheck_skills, doublecheck_spec, and doublecheck_report tools, and the per-dimension verification workflow. No consensus, no implementation.
• Discipline guard (doublecheck-guard) — the grill gate, the red/green evidence gates, the adversary review, the /doublecheck and /gate commands, the live gate settings card, and the four-phase delivery gate, at remind / warn / block intensity.
• Evidence from the log, not from self-reports — every injected reminder, review, and gate notice rides the standard channels into the session log; durable facts ride doublecheck/state, doublecheck/gate, or tool results. Replay is the state, so a resumed or forked session re-derives the same /gate run result.
• One failed check, one red light — a missing spec, a latest run that is not green, coverage below the threshold, an edit unmapped to any spec dimension, or a blocker/major finding each becomes a red light with a rework suggestion; warnings and skips never flip the verdict.
• Audit-safe — gate and delivery reports record counts, ids, and verdicts only, with no file contents or session text; model-produced finding texts pass a secret redactor before storage or display.

The loop is one line: grill → design → red → green → review → verify. Each step leaves replayable evidence, while the six requirement dimensions, the consensus gate, and the structured spec land in both the session and the workspace.

The six stages of the discipline loop:

Stage	Meaning
grill	Interrogate the six requirement dimensions; refuse to implement until consensus.
design	The settled spec is committed via doublecheck_spec.
red	A failing test run proves the gap before implementation edits.
green	A passing test run after the edits closes the loop.
review	A forked adversary critic audits the delivery against the spec.
verify	doublecheck_report plus a per-dimension verification workflow prove the delivery.

Tools & surfaces

Surface	Kind	Notes
doublecheck_skills	tool	Lists and loads the package's four bundled skills through the skill registry seam.
doublecheck_spec	tool	Commits the grilled six-dimension spec to the session log and a workspace markdown copy.
doublecheck_report	tool	Folds the discipline evidence into a delivery report (optional per-dimension verification workflow).
/doublecheck status|report|on|off	command	Switch, modules, intensity, stage facts, folded report, and the durable on/off override.
/gate status|run|config	command	Live checklist progress, the settled deliverable/rework report, and the effective config.
grill-requirements, red-green-tdd, delivery-review, delivery-proof	skill	Bundled discipline skills covering all six loop stages.
gate	live config field	The pluggable checklist is the guard row's one .volatile() field: it is edited from the row's settings card (namespace = the row's profile entry id) and read once at load.
strict.patch.yml	overlay	Every gate on at block intensity plus the coverage requirement, in one patch layer.
dsh-doublecheck/invariant	companion row	Reports package-owned write-path contradictions through the host invariants registry.
doublecheck	session projection	One bounded status line a client renders: stage, latest gate verdict, red-light count, edits, spec/review flags. See Reading the status line.

Reading the status line

The doublecheck session projection is a plain string, and it is deliberately one line — the whole point is that a client can show it without cutting anything off. A chat bridge's /status, a status bar, or any custom consumer renders it as it is:

gate:rework r2 | stage:green | test:green edits:12 spec:yes review:yes | goal:ship parser fix

Segment	Reads as
gate:<verdict> r<n>	The latest delivery-gate verdict — none before any run, deliverable, or rework — and its red-light count r<n>.
stage:<stage>	The discipline stage reached on durable evidence: grill, design, red, green, review, verify.
test:<color>	The latest test-run color: none, red, green.
edits:<n>	Implementation edits folded from the session log (test-file writes are not counted).
spec:yes / spec:no	Whether a committed doublecheck_spec exists.
review:yes / review:no	Whether an adversary review record exists.
goal:<excerpt>	The committed spec goal, clipped. Present only when spec:yes.

The line is bounded by construction, so no renderer truncates it:

• Hard budget: 120 characters. A renderer that keeps values up to 120 characters — the WeChat bridge's generic projection renderer is one — shows the entire line. The old object-shaped payload was 131 characters when empty, which is why the verdict and the red count used to be cut off.
• The first three segments are lossless. Counts above 9999 red lights or 99999 edits print as 9999+ / 99999+ instead of growing without bound; a non-finite or negative count prints as 0.
• goal: is the one lossy segment. It is flattened to a single line (newlines and control characters become spaces, | becomes /) and clipped to the budget that is left, at most 40 characters, ending in … when it was clipped. The complete goal is in /doublecheck status and the delivery report.
• This is a glance, not the record: /gate status and /doublecheck status render the full state.

The projection's stateVersion is 3. Bumping it discards a persisted projection-cache row written by an older version and refolds the key from the session log, so an upgraded session recomputes the line instead of showing a stale one. The persisted fold state itself is unchanged, and a pre-upgrade session never fails to load.

Gate phases

The delivery gate aggregates the session's durable evidence into a configurable four-phase checklist and settles one deliverable / rework required decision. Every phase folds the session log alone (replay IS the state), so a run re-derives identically after resume or fork.

Phase	Checks	Evidence source	Model cost
Requirements interrogation	Key-question checklist confirmed item by item (six spec-dimension questions by default)	Committed doublecheck_spec + ask_user_question calls	none
Test evidence	Latest run color, failing runs after green, optional coverage threshold, optional dsh-eval report	Shell test runs in the session log ([exit code: N], coverage percentages); the dsh-eval report file when gate.tests.evalReports.enabled	none
Implementation consistency	Diff ↔ requirement mapping: every edit must serve a spec dimension	Local forked reviewer (structured findings, read-only tools)	one subagent
Review conclusion	The delivery verdict; engine: auto consumes dsh-auto-review's durable verdict records when present, else the local reviewer	autoReview/verdict / autoReview/rejection events, or the local forked reviewer	one subagent (local)

Red lights are failed checks (a missing spec, a failing latest run, coverage below minimum, an unmapped edit, blocker/major findings) — each carries a rework suggestion. Warnings and skips never flip the decision. The gate integrates dsh-auto-review as a weak dependency: review.engine: auto folds its verdict records when present and degrades to the local reviewer otherwise; gate.tests.evalReports.enabled folds its eval engine's dsh-eval report (prompt-regression / stress / fairness suites) into the test evidence and skips honestly when no report exists. The gate never synthesizes approval requests.

Examples

These are the plugin's real entry points; install once, then call one.

Detailed example

Command

/gate run

Output

# Delivery gate report

> **Verdict: rework required** — 2 red item(s)
> The gate is red. Re-open the work in plan mode to re-check the open items before delivering.

Observed from README.md §Example report (lines 243-251)

Failure branch

Input

/doublecheck maybe

Output

Unknown /doublecheck argument "maybe". Usage: /doublecheck status|report|on|off

Observed from plugins/dsh-doublecheck/src/guard/prose.ts:191-192; plugins/dsh-doublecheck/src/guard/command.ts:156; plugins/dsh-doublecheck/tests/command.spec.ts:201-208

Brief examples

• doublecheck_skills
• doublecheck_spec

Example report

/gate run returns this markdown — paste it into a PR description:

# Delivery gate report

> **Verdict: rework required** — 2 red item(s)
> The gate is red. Re-open the work in plan mode to re-check the open items before delivering.

## 1. Requirements interrogation — PASS
- [✔] **What outcome must the delivery produce?** — spec dimension "goal" committed
- [✔] **What is in scope, and what is out of scope?** — spec dimension "scope" committed
- [✔] **Which observable checks prove the work is done?** — spec dimension "acceptanceCriteria" committed
- [✔] **What can go wrong, and what is the correct behavior in each case?** — spec dimension "failureModes" committed
- [✔] **What is traded when goals conflict; what is optional?** — spec dimension "priorities" committed
- [✔] **What does the user explicitly not want?** — spec dimension "nonGoals" committed

## 2. Test evidence — FAIL
- [✔] **passing test run** — latest test run passed
- [✔] **failing cases after green** — 0 failing run(s) after green (allowed: 0)
- [✖] **coverage evidence** — 61% coverage below the 80% minimum — rework: raise coverage above the configured minimum

## 3. Implementation consistency — WARN
- [⚠] **[minor] src/telemetry.ts touched without a requirement** — [minor] the edit adds a metric no spec dimension covers

## 4. Review conclusion — PASS
- [✔] **dsh-auto-review conclusion** — 3 call(s) approved by dsh-auto-review (latest risk: low)

## Red items
1. **tests/coverage** — 61% coverage below the 80% minimum — *rework: raise coverage above the configured minimum*
2. **consistency/finding-1** — [minor] the edit adds a metric no spec dimension covers — *rework: src/telemetry.ts touched without a requirement*

## Audit
- review engine: dsh-auto-review
- generated at: 2026-08-14T12:00:00.000Z
- counts, ids, and verdicts only: no file contents or session text are embedded, and recognized secrets are redacted.

CI output

/gate run also writes a gate-report.json (the same settled state as lossless JSON, next to gate-report.md). The doublecheck-gate CLI turns that file into machine-readable output for GitHub Actions:

# JSON (PR comment / status payload)
doublecheck-gate --format json --input gate-report.json
# SARIF 2.1.0 (code-scanning upload / status check)
doublecheck-gate --format sarif < gate-report.json

The CLI only serializes the already-settled GateState — it never re-runs the four-phase gate or the evidence folds. Its exit code maps the verdict: 0 = deliverable, 1 = rework, 2 = usage/parse error.

Configuration

All tunables are Schemastery Config fields (changeable from cordis.yml). An id-targeted override replaces the whole row — restate every key you need. cordis.patch.yml documents each key inline; Schema defaults are the single source of tuning defaults.

Most-used keys:

Key	Default	Meaning
specFile	'doublecheck-spec.md'	Workspace file for the committed spec markdown (grill row).
reportFile	'doublecheck-report.md'	Workspace file for the delivery report (grill row).
intensity	'remind'	Enforcement strength of the grill, red/green, and review gates (remind / warn / block).
enableByDefault	true	Master switch for sessions without a /doublecheck on|off record.
language	'en'	Injected reminder/deny/review/gate prose language (en / zh).
modules.adversary	false	On enables the forked critic review at green.
gate.enabled	true	Master switch for the gate panel and the turn-boundary red notice.
gate.review.engine	'auto'	auto = dsh-auto-review verdict records when present, else the local reviewer; local = always local.

Key	Default	Meaning
specFile	'doublecheck-spec.md'	Workspace file for the committed spec markdown (grill row).
reportFile	'doublecheck-report.md'	Workspace file for the delivery report (grill row).
reportVerify	true	Run the verification workflow by default (grill row).
verifyProvider	'fork'	Subagent provider for the per-dimension checkers (grill row).
verifyMode	'all'	all = one parallel checker per dimension; single = one combined checker (grill row).
intensity	'remind'	Enforcement strength of the grill, red/green, and review gates (remind / warn / block).
enableByDefault	true	Master switch for sessions without a /doublecheck on|off record.
language	'en'	Injected reminder/deny/review/gate prose language (en / zh).
guardTools	['edit', 'write']	Mutation tool names both gates watch.
vagueTaskMaxChars	200	Longer tasks are never treated as vague.
remindOnce	true	Inject each reminder at most once per session (durable across restarts).
testToolNames	['bash', 'pwsh']	Shell tool names that can run tests.
testCommandPatterns	(pnpm / npm / yarn / bun test, pytest, go / cargo / make test, node --test, deno test, uv run pytest)	Regexes a command must match to count as a test run.
testFilePatterns	(test dirs, *.test.* / *.spec.*)	Regexes identifying test files — always editable, exempt from the red gate.
modules.grill	true	Off disables the grill gate.
modules.tdd	true	On enables the red/green evidence gates.
modules.adversary	false	On enables the forked critic review at green.
adversaryModel	null	Critic model route; null = main model self-reviews.
adversaryProvider	'fork'	Subagent provider the critic runs on.
adversaryMaxFindings	5	Findings cap (1–20) injected into the session.
adversaryTools	['read', 'glob', 'grep']	Critic tool allowlist; keep it read-only.
adversaryTimeoutMs	120000	Hard time budget for one critic run.
gate.enabled	true	Master switch for the gate panel and the turn-boundary red notice.
gate.planSuggestion	true	Append the plan-mode re-check suggestion to red reports.
gate.reportFile	'gate-report.md'	Workspace file for the gate report.
gate.requirements.checklist	(six spec-dimension questions)	Pluggable key-question checklist: { id, question, specDimension, required }.
gate.requirements.minConfirmed	6	Minimum required questions that must pass (1..required count).
gate.requirements.interrogateTool	'ask_user_question'	Tool name whose calls count as interrogation evidence.
gate.tests.requirePassingRun	true	A non-passing (or missing) latest test run is a red light.
gate.tests.allowFailingRuns	0	Failing runs after the latest green allowed before red.
gate.tests.requireCoverage	false	On requires coverage evidence in the test output.
gate.tests.minCoveragePct	80	Minimum coverage percentage (0–100).
gate.tests.evalReports.enabled	false	On folds the dsh-eval report (dsh-auto-review's eval engine) into the test evidence.
gate.tests.evalReports.dir	'.eval-reports'	Workspace-relative directory holding the engine's report.
gate.tests.evalReports.file	'report.json'	Report file name inside the directory.
gate.tests.evalReports.required	false	A missing report is a red light exactly when true (a skip otherwise).
gate.consistency.*	provider: 'fork', model: null, tools: ['read','glob','grep'], timeoutMs: 120000, maxFindings: 5	The local consistency reviewer's knobs (model: null = main model).
gate.review.engine	'auto'	auto = dsh-auto-review verdict records when present, else the local reviewer; local = always local.
gate.review.provider	'fork'	The local review reviewer's provider (its model/tools/timeoutMs/maxFindings match gate.consistency.*).

Every gate.* key above is one live field: the guard row marks the whole gate block .volatile(), so it is the row's editable settings card (the form namespace is the row's profile entry id). The remaining keys are ordinary composition config from the profile patch. The block is read once at load, so an edit takes effect on the next load — gate.enabled decides whether the turn-boundary red notice is installed at all. On a host line whose schemastery predates .volatile() the row still mounts: the card is absent and the block comes from the profile patch exactly as before.

Misconfiguration fails loud at load: invalid regexes, empty or duplicated name lists, out-of-range thresholds, and duplicate checklist ids throw instead of silently doing nothing. strict.patch.yml is the all-gates-block overlay that restates the guard row at intensity: block with every module on and the coverage requirement enabled.

Example override in your profile patch:

- insert:
    - id: doublecheck-guard
      name: 'dsh-doublecheck/guard'
      config:
        language: zh
        intensity: block
        modules:
          tdd: true
          adversary: true
        gate:
          tests:
            requireCoverage: true

FAQ

Durable facts are two session events, doublecheck/state (the durable switch) and doublecheck/gate (the settled gate state), plus four workspace files — doublecheck-spec.md, doublecheck-report.md, gate-report.md, gate-report.json (paths configurable). Gate and delivery reports record counts, ids, and verdicts only: no file contents and no session text, and model-produced finding texts pass a secret redactor before storage or display. Nothing outside the session workspace is touched.

Nothing is silent. The durable records behind /doublecheck on|off and /gate run ride the host's ignorable append surface; on a host without it the writes are skipped and the switch falls back to process-local state (see Known limitations). The guard row's settings card needs the settings service: where the host's schemastery predates .volatile() there is no card, but the row still mounts and the gate block still comes from the profile patch. When a reviewer or adversary seam cannot run, the conclusion is written honestly as "unavailable"/skip instead of a fabricated verdict.

No. Every tool is namespaced, it provides no service key, registers no client slot and no webServer prefix, and every shared-event listener delegates through next() and never short-circuits the chain. The point-by-point argument is in "Interoperability with other DSH plugins".

The requirements and test-evidence phases are pure folds with no model cost; the consistency and review-conclusion phases start one forked subagent each (with gate.review.engine: auto and dsh-auto-review verdict records present, the review-conclusion phase can be consumed at zero cost). The optional adversary review adds one more. Nothing else calls a model or the network.

Security boundaries

• Model-visible ⟺ logged. Every injected reminder, review, and gate notice rides the standard channels and lands in the session log; the durable spec/state/gate facts ride tool results or SessionEventMap members.
• Fail closed / fail loud. Guard and gate config are validated in apply (assertions throw); a reviewer or adversary seam that cannot run settles as an honest "unavailable"/skip notice instead of a fake verdict.
• Audit-safe reports. Gate and delivery reports record counts, ids, and verdicts only — no file contents or session text — and model-produced finding texts pass a secret redactor before storage or display.
• No network of its own. The plugin makes no direct network requests; the critic and reviewer subagents ride the harness subagent seam.
• Weak dependency on dsh-auto-review. It is never imported or hard-required; the gate folds its durable verdict records and degrades to the local reviewer, and never synthesizes approval requests.

Permissions & data

• Reads: the session log (tool/call / tool/result / tool/ptc-dispatch, injected user/message sources, and the foreign autoReview/* verdict records) in-process only; the optional plan-mode service state. PTC sub-dispatches carry the predecessor tool/code-dispatch label on hosts before the V3 rename; both labels fold identically. Injected notices carry the producer-owned dsh-doublecheck message-source kind; logs recorded before that kind existed arrive as plugin:dsh-doublecheck (the harness's V3→V4 migration on a released catch-all wrapper) or as the released plugin wrapper, and all three fold identically.
• Writes: doublecheck-spec.md, doublecheck-report.md, gate-report.md, and gate-report.json in the session workspace (paths configurable) through the ctx.fs seam; the durable doublecheck/state and doublecheck/gate session events.
• Model calls: the gate's consistency and local-review phases (one subagent each per /gate run), the optional adversary review, and the doublecheck_report verification workflow start subagent runs; nothing else calls a model or the network.
• Never touched: credentials, environment variables, or any file outside the session workspace. The workshop manifest declares filesystem:read and filesystem:write only. Gate reports carry counts, ids, and verdicts only; recognized secrets in reviewer texts are redacted before storage or display.

Known limitations

• Durable writes. /doublecheck on\|off → doublecheck/state and /gate run → doublecheck/gate ride the host's ignorable append surface (post-rc.6 through 0.1.1-rc.2). On hosts without that surface (rc.6/rc.8, and 0.1.2-alpha.1, which removed the envelope — 0.1.2-rc.1 restores the field for stored-log read compatibility only and still cannot stamp it), the writes are skipped and the switch stays process-local. 0.1.2-rc.1 (adapted 2026-09-02): the session envelope keeps its ignorable field for stored-log read compatibility only - Session.append still cannot stamp it, so audit-gate behavior is unchanged. 0.1.5-alpha.1 (adapted 2026-09-09): session format V3 renames the durable sub-dispatch event tool/code-dispatch to tool/ptc-dispatch (payload unchanged; both labels fold identically). Session.append still exposes no ignorable channel, so durable writes stay skipped and the switch stays process-local - behavior unchanged. (The doublecheck-gate settings namespace this entry introduced was removed in 0.1.7-alpha.1, when the harness deleted the namespace registry.) 0.1.5-rc.1 (adapted 2026-09-10): dependency pins move to the published 0.1.5-rc.1 line; no seam change affects this plugin's behavior. 0.1.5-rc.2 (adapted 2026-09-11): dependency pins move to the published 0.1.5-rc.2 line; no seam change affects this plugin's behavior. 0.1.7-alpha.1 (adapted 2026-10-04): the harness deleted the shared catch-all plugin message-source kind (notices now carry the producer-owned dsh-doublecheck kind, and the remindOnce fold still reads the two pre-upgrade shapes a durable log can hold), and it replaced the settings-namespace registry with SettingsForms (the doublecheck-gate namespace is gone; the gate checklist is now the guard row's one .volatile() field, still read once at load). Dev/test pins for @deepseek-ai/cordis and @deepseek-ai/schemastery move to ^4.0.3 / ^3.18.3, where Volatile and .volatile() first exist; the peer ranges stay ^4.0.2 / ^3.18.2 and the live-field surface is detected at load, so the older host lines mount the row with no settings card. No behavior change beyond the settings storage location.
• Optional seams. The guard row's settings card appears when the settings service is mounted; the card is this row's own Config (namespace = its profile entry id), and the gate block is read once at load, so its values apply to the /gate panel and the gate-red notice on the next load. The live-field surface is detected at load, so on a host line whose schemastery predates .volatile() the row mounts anyway — without a card, with the gate block taken from the profile patch. The /gate status plan-mode line reads the optional ctx.planMode (shows unknown without it); the adversary review needs ctx.subagents; verification needs workflowEngine.
• Local degrade. gate.review.engine: auto degrades to the local reviewer when dsh-auto-review is absent or has no verdict records this session — the report names the reason instead of inventing a verdict.
• dsh-eval evidence is file-based. The dsh-auto-review eval engine (dsh-eval) writes its prompt-regression / stress / fairness results to a workspace report file, not the session log. gate.tests.evalReports.enabled folds that file (off by default; skips when absent) and the folded counts ride the durable doublecheck/gate record so a settled run still replays.

Interoperability with other DSH plugins

Verified against DSH 0.2.0-rc.2 (the runtime this README ships for) and the high-star plugin set surveyed on 2026-10-05.

This plugin does not interfere with other plugins, including the widely installed high-star ones:

• No tool-name collision. Every tool is namespaced; no bare name owned by a shipped tool or another plugin is registered.
• No service-key collision. It provides no service key at all, so it cannot collide on one.
• No slot collision. It registers no client slot key, so it cannot contend for a shadows-shipped-ui seat.
• No HTTP route collision. It registers no webServer prefix.
• No patch-layer collision. The bundle patch only inserts its own row; it never overrides a built-in row's config.
• No global mutation. It does not patch prototypes, rewrite process.env, or replace the global fetch dispatcher.

Shared event listeners are non-interfering by construction. It observes the ordering-sensitive events agent/turn-stopping, tools/post-execute, tools/pre-execute with ctx.on() — Cordis's broadcast registration, where every listener runs and none can starve another. Every listener here delegates through next(), so the chain is never short-circuited, and a mutation is applied to the value next() produced rather than returned in its place:

• tools/post-execute — also used by cc-safety-net (1576★).
• tools/pre-execute — also used by cc-safety-net (1576★).

Static evidence: dsh-plugin-doctor K10–K13 report pass for every check on this repository.

Compatibility

Surface	Status
Harness	DeepSeek Harness dsh-v0.2.2-rc.1 (adapted 2026-11-05): engines.dsh gains >=0.2.2-0 <0.3.0 (every earlier line stays admitted); the gate and projection paths are covered by the rc.1 test suite.
Node	^22.19.0 || >=24.9.0
Platforms	All (pure host; no native code, no direct network requests of its own)
Model	Any (the guard itself never calls a model; the critic and reviewer phases run as harness subagents)

Ecosystem & quality

• Gitee mirror (sync mirror; GitHub is canonical)
• dshfind plugin page (installs and ranking)
• OpenSSF Scorecard report (driven by .github/workflows/scorecard.yml)
• dsh-plugin-doctor verification report (K10–K13 all pass)
• DSH Market (listed)
• 1024 store channel: npm i -g dsh1024 once, then dsh1024 plugin --profile web add dsh-doublecheck (counts toward the deepseek1024.com install ranking)
• Ecosystem knowledge base — measured data, not marketing: plugin development guide · plugin-selection data · maintenance criteria

Development

pnpm install             # node ^22.19 || >=24
pnpm run build           # tsc --noEmitOnError (lib/ is committed)
pnpm run prepare         # tsc --noEmitOnError (git-install channel)
pnpm run prepublishOnly  # build + full test suite
pnpm run typecheck       # tsc --noEmit + tests tsconfig
pnpm run lint            # eslint src tests
pnpm test                # vitest run
pnpm run test:coverage   # vitest run --coverage
pnpm run pack:check      # build + pack the tarball
pnpm run check:readmes   # five-language README consistency gate

lib/ is committed on purpose: the git-install channel resolves the package without a build step, so rebuild with pnpm run build and commit the regenerated tree after source changes. The README images are generated programmatically by scripts/gen-readme-assets.py: Pillow draws the hero banner (zh/en × light/dark WebP) and hand-written SVG draws the structure diagram (flow-v5, light/dark), so the script is safe to re-run and overwrites idempotently. The older single-skin flow diagram and demo images stay in docs/assets/ for external links; the READMEs reference only the new hero and flow-v5.

Topics: dsh, dsh-plugin, deepseek-harness, engineering-discipline, requirements, guard, skill, quality-gate, delivery-gate

Family & thanks

This project is part of the DeepSeek Harness plugin family from PerryLink (all Apache-2.0). If it is useful, a star helps the next person find it — nothing is gated behind it.

• @PerryLink — creator and maintainer: the grill → design → red → green → review → verify discipline loop, the four-phase delivery gate, the five-language docs, and the CI/release pipeline.

Plugin	Status
dsh-aqua-input-check
dsh-archive-check
dsh-auto-review
dsh-autotier	🧊 FROZEN — see the repo README
dsh-bid-ca-precheck
dsh-bid-qual-check
dsh-cert-mcp	🧊 FROZEN — see the repo README
dsh-checkpoint-rewind
dsh-click
dsh-composer-history
dsh-contract-stance
dsh-customs-doc-check
dsh-data-quality	🧊 FROZEN — see the repo README
dsh-demurrage-ledger
dsh-doublecheck
dsh-drill-script-check
dsh-eia-guide-check
dsh-emergency-plan
dsh-essay-rubric-check
dsh-evidence-check
dsh-export-ctl-check
dsh-fast
dsh-fmea-table-check
dsh-forecast-penalty
dsh-fund-research
dsh-github
dsh-gongwen-flow-check
dsh-gongwen-word-check
dsh-guard-plan-qc
dsh-hazchem-check
dsh-hazplan-check
dsh-hidden-risk-map
dsh-hs-classify
dsh-icd-rule-check
dsh-industry-research
dsh-lawcite-adapter
dsh-laya
dsh-lc-doc-check
dsh-learning-gap-check
dsh-library
dsh-local-ai
dsh-lsp-actions
dsh-mask
dsh-mcp-panel
dsh-medrec-qc
dsh-memento	🧊 FROZEN — see the repo README
dsh-nsfc-form-check
dsh-nurse-record-check
dsh-observe
dsh-origin-rvc-check
dsh-ota-review-check
dsh-output-styles
dsh-paper-doc-adapter
dsh-permission-rules
dsh-permit-report-check
dsh-pipeline-check
dsh-pleading-draft
dsh-plugin-doctor	🚫 RETIRED (repo archived)
dsh-plugin-guide
dsh-plugin-kit	🚫 RETIRED (repo archived)
dsh-plugin-upgrade	🚫 RETIRED (repo archived)
dsh-policy-brief-draft
dsh-power-loss-split
dsh-power-ticket-check
dsh-ppap-check
dsh-protest-deadline
dsh-railway-window
dsh-repair-order-qc
dsh-research-report
dsh-review-reply-check
dsh-rulefile-check
dsh-safety-brief-check
dsh-score	🧊 FROZEN — see the repo README
dsh-session-sync
dsh-site-log-check
dsh-skill-pack-security
dsh-soe-decision-check
dsh-soilwater-check
dsh-sop-sync-check
dsh-spc-gbt-adapter
dsh-talk
dsh-tender-extract
dsh-tender-matrix
dsh-test-drive
dsh-ticktick	🚫 RETIRED (repo archived)
dsh-translate
dsh-warranty-calc
dsh-budget	🧊 FROZEN — see the repo README
dsh-claude-move	🧊 FROZEN — see the repo README
dsh-defend	🧊 FROZEN — see the repo README
dsh-draw	🧊 FROZEN — see the repo README
dsh-reach	🧊 FROZEN — see the repo README
dsh-background-agents	🚫 RETIRED (repo archived)
dsh-session-pin	🚫 RETIRED (repo archived)
dsh-team-rooms	🚫 RETIRED (repo archived)

License

Apache License 2.0 © 2026 dsh-doublecheck contributors