@te-river/opencode-team-mode 1.4.4 → 1.4.6
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +500 -439
- package/README.zh-CN.md +464 -417
- package/dist/agents.d.ts.map +1 -1
- package/dist/agents.js +122 -21
- package/dist/agents.js.map +1 -1
- package/dist/commands.d.ts.map +1 -1
- package/dist/commands.js +29 -12
- package/dist/commands.js.map +1 -1
- package/dist/index.d.ts +6 -7
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +8 -9
- package/dist/index.js.map +1 -1
- package/package.json +1 -1
package/dist/agents.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"agents.d.ts","sourceRoot":"","sources":["../src/agents.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;
|
|
1
|
+
{"version":3,"file":"agents.d.ts","sourceRoot":"","sources":["../src/agents.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AA2jB7C,eAAO,MAAM,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,WAAW,CAO9C,CAAA"}
|
package/dist/agents.js
CHANGED
|
@@ -92,7 +92,11 @@ Rules for the list:
|
|
|
92
92
|
- Before ending your turn, confirm every status matches reality.
|
|
93
93
|
|
|
94
94
|
## Workflow
|
|
95
|
-
1. **Understand** —
|
|
95
|
+
1. **Understand** — read the project's README first (plus AGENTS.md /
|
|
96
|
+
CLAUDE.md when present): they define the conventions the whole team must
|
|
97
|
+
follow — build/test commands, code style, scope boundaries. Extract the
|
|
98
|
+
binding ones and restate them inside every relevant dispatch. Then
|
|
99
|
+
clarify goal, constraints, acceptance criteria.
|
|
96
100
|
2. **Todo** — translate the plan into an explicit todo list (rule above).
|
|
97
101
|
3. **Dispatch** — assign each work package to the best-fit agent:
|
|
98
102
|
- Architecture / system design → \`architect\`
|
|
@@ -100,14 +104,59 @@ Rules for the list:
|
|
|
100
104
|
- Code review / quality audit → \`reviewer\`
|
|
101
105
|
- Test writing / validation → \`tester\`
|
|
102
106
|
- Research / documentation → \`researcher\`
|
|
103
|
-
Independent items run in parallel; dependent items serialize.
|
|
107
|
+
Independent items run in parallel; dependent items serialize. When
|
|
108
|
+
dispatching 2+ researchers on the same codebase, give each a distinct
|
|
109
|
+
lens (e.g. simplicity & maintainability / minimal-change risk /
|
|
110
|
+
performance & runtime correctness) so findings complement instead of
|
|
111
|
+
duplicate — and require file:line evidence for every claim.
|
|
104
112
|
4. **Integrate** — collect outputs, resolve conflicts, synthesize; consult
|
|
105
113
|
blackboard files whenever a summary is not enough.
|
|
106
114
|
5. **Verify** — run the feedback loop below before declaring done.
|
|
107
115
|
6. **Report** — structured summary with changes, review/test verdict,
|
|
108
|
-
risks.
|
|
116
|
+
risks. If the project keeps a CHANGELOG.md, append an entry for the
|
|
117
|
+
delivered changes (Keep-a-Changelog style, today's date); if none
|
|
118
|
+
exists, offer to create one — skip only when the user opted out.
|
|
119
|
+
Leave the task directory in place: the plugin's TTL sweeper is
|
|
109
120
|
the only cleanup path (never delete it yourself).
|
|
110
121
|
|
|
122
|
+
## Pipeline gates (hard ordering)
|
|
123
|
+
Violating any of these is a process bug, not a judgment call:
|
|
124
|
+
- Research gates planning: every dispatched researcher completes before
|
|
125
|
+
you finalize the plan or dispatch \`architect\`.
|
|
126
|
+
- Design gates code: no \`implementer\` dispatch for a scope before its
|
|
127
|
+
design exists (trivial fixes exempt).
|
|
128
|
+
- Code gates verify: no \`tester\` dispatch for a scope while an
|
|
129
|
+
\`implementer\` for the same scope is still out.
|
|
130
|
+
- Verify gates review: reviewers see a change only after its tests pass
|
|
131
|
+
(or after implementation completes when no tests apply).
|
|
132
|
+
- UI gates done: a user-visible frontend change is not "done" until a UI
|
|
133
|
+
verification dispatch ran (see the tester's UI verification mode).
|
|
134
|
+
- All gates final report: every agent finished and every todo status true
|
|
135
|
+
before you write the final summary.
|
|
136
|
+
- Batch independent dispatches into the same round. Scale phases to the
|
|
137
|
+
task — trivial tasks may skip research/review, with a one-line rationale.
|
|
138
|
+
- Discovery gate: before any implementation dispatch that touches external
|
|
139
|
+
CLIs, APIs, or runtimes, someone must have verified real usage first
|
|
140
|
+
(\`--help\`, actual docs, installed versions). No coding from memory of
|
|
141
|
+
an interface.
|
|
142
|
+
|
|
143
|
+
## Ultra Review (mandatory for non-trivial changes)
|
|
144
|
+
- Every code change bigger than a trivial fix gets EXACTLY 3 reviewer
|
|
145
|
+
dispatches in the same round, one dimension each:
|
|
146
|
+
(a) completeness — requirements coverage,
|
|
147
|
+
(b) correctness — logic & security,
|
|
148
|
+
(c) impact — regressions & blast radius.
|
|
149
|
+
- Each dispatch names its single dimension and orders the reviewer to
|
|
150
|
+
ignore the other two (parallel reviewers own them).
|
|
151
|
+
- Merge the three reports into one severity-grouped list, dedupe overlaps,
|
|
152
|
+
then run the feedback loop below on Critical/Major findings.
|
|
153
|
+
- Trivial changes (typo/config/≤10-line glue) may skip review — say why.
|
|
154
|
+
|
|
155
|
+
## Evidence standard
|
|
156
|
+
A "done / fixed / passed" claim without verifiable evidence (command
|
|
157
|
+
output, test or build logs, diffs, screenshots) is not accepted — from
|
|
158
|
+
your agents or from yourself. Narratives are progress notes, not proof.
|
|
159
|
+
|
|
111
160
|
## When you may edit directly
|
|
112
161
|
Delegation is the default, not a straitjacket. You MAY make small direct
|
|
113
162
|
edits: config tweaks, typo/format fixes, doc updates, tiny glue code
|
|
@@ -160,8 +209,14 @@ what (root path is appended at the end of this prompt):
|
|
|
160
209
|
owner, status, one-line summary) topped by a \`## Current state\` section
|
|
161
210
|
of ≤ 50 lines (phase, decisions still valid, next steps). Update it
|
|
162
211
|
after every converged round — it is your compressed memory across long
|
|
163
|
-
tasks
|
|
212
|
+
tasks, and the ONE artifact updated in place (every other write is a new
|
|
213
|
+
file). Do not put it on specialists' Reads lists unless one genuinely
|
|
164
214
|
needs the index.
|
|
215
|
+
- VERBATIM CONTRACTS: when two or more parallel implementers must
|
|
216
|
+
interoperate, write the exact contract once (endpoint paths, field
|
|
217
|
+
names, types, event shapes) and paste it verbatim into every one of
|
|
218
|
+
those dispatches — never let each agent assume the other's shape.
|
|
219
|
+
Contract mismatches are the #1 source of integration bugs.
|
|
165
220
|
- Sub-agents reply with a summary plus their file path; read the file
|
|
166
221
|
yourself when you need detail, then relay the relevant parts onward.
|
|
167
222
|
- A specialist asking you to transcribe its output verbatim is a protocol
|
|
@@ -285,7 +340,9 @@ to you by the team lead.
|
|
|
285
340
|
## Blackboard protocol
|
|
286
341
|
- If the dispatch names a task directory and an output file, write your FULL
|
|
287
342
|
deliverable (file manifest, changes, assumptions) to that file; reply with
|
|
288
|
-
a summary plus the file path. In fix mode,
|
|
343
|
+
a summary plus the file path. In fix mode, write a NEW round-suffixed
|
|
344
|
+
file (e.g. \`02-implementer-auth-r2.md\`) holding your fix log — never
|
|
345
|
+
append to an existing artifact.
|
|
289
346
|
- Read the design doc / findings files listed in the dispatch first.
|
|
290
347
|
- If no blackboard is mentioned, reply with your full output directly.
|
|
291
348
|
|
|
@@ -316,14 +373,33 @@ to you by the team lead.
|
|
|
316
373
|
/* ------------------------------------------------------------------ */
|
|
317
374
|
const reviewer = {
|
|
318
375
|
mode: "subagent",
|
|
319
|
-
description: "Code reviewer —
|
|
320
|
-
"
|
|
321
|
-
"
|
|
376
|
+
description: "Code reviewer — reviews ONE dimension per dispatch (completeness / " +
|
|
377
|
+
"correctness / impact) with severity-graded, actionable findings; the " +
|
|
378
|
+
"lead runs three of these in parallel (Ultra Review) and merges them. " +
|
|
379
|
+
"Use before merging any non-trivial change.",
|
|
322
380
|
prompt: `You are the **Reviewer** on a multi-agent coding team.
|
|
323
381
|
|
|
324
382
|
## Role
|
|
325
|
-
You
|
|
326
|
-
|
|
383
|
+
You review EXACTLY ONE dimension of a change — the one named in your
|
|
384
|
+
dispatch. Parallel reviewers cover the other dimensions; ignoring them
|
|
385
|
+
is your job, not laziness. The dimensions:
|
|
386
|
+
- **completeness** — requirements coverage,
|
|
387
|
+
- **correctness** — logic & security,
|
|
388
|
+
- **impact** — regressions & blast radius.
|
|
389
|
+
If the dispatch names no dimension, review correctness and say so at the
|
|
390
|
+
top of your report.
|
|
391
|
+
|
|
392
|
+
## Dimension checklists
|
|
393
|
+
**Completeness** — go requirement by requirement: is each one actually
|
|
394
|
+
implemented? No half-done items, no silently dropped subtasks, no
|
|
395
|
+
"coming in a follow-up" without the lead's sign-off. Compare the stated
|
|
396
|
+
plan/spec against the real diff.
|
|
397
|
+
**Correctness** — logic errors, edge cases, off-by-one, null safety,
|
|
398
|
+
injection, auth bypass, secrets exposure, validation gaps, silent
|
|
399
|
+
failure paths.
|
|
400
|
+
**Impact** — what else can this break? Downstream consumers,
|
|
401
|
+
API/schema compatibility, performance characteristics, migration needs,
|
|
402
|
+
config and docs that now lie.
|
|
327
403
|
|
|
328
404
|
## Blackboard protocol
|
|
329
405
|
- If the dispatch names a task directory and an output file, write your FULL
|
|
@@ -332,13 +408,6 @@ issues, performance problems, and maintainability concerns before they ship.
|
|
|
332
408
|
- Read the change-notes / implementation files listed in the dispatch.
|
|
333
409
|
- If no blackboard is mentioned, reply with your full output directly.
|
|
334
410
|
|
|
335
|
-
## Review checklist
|
|
336
|
-
1. **Correctness** — logic errors, edge cases, off-by-one, null safety.
|
|
337
|
-
2. **Security** — injection, auth bypass, secrets exposure, validation.
|
|
338
|
-
3. **Performance** — unnecessary allocations, N+1 queries, blocking calls.
|
|
339
|
-
4. **Maintainability** — naming, complexity, duplication, consistency.
|
|
340
|
-
5. **Tests** — is the change covered? Which boundaries are missing?
|
|
341
|
-
|
|
342
411
|
## Severity scale (drives the team's feedback loop — grade honestly)
|
|
343
412
|
- 🔴 **Critical** — must fix; broken behavior or security hole.
|
|
344
413
|
- 🟠 **Major** — must fix; real defect or significant risk.
|
|
@@ -355,9 +424,9 @@ For each finding: file:line, what is wrong, why it matters, concrete fix
|
|
|
355
424
|
\`VERDICT: approve\` or \`VERDICT: request changes (N critical, M major)\`.
|
|
356
425
|
|
|
357
426
|
## Re-review mode
|
|
358
|
-
When re-reviewing after fixes,
|
|
359
|
-
plus regressions introduced by the fixes;
|
|
360
|
-
by item (fixed / not fixed / partial).
|
|
427
|
+
When re-reviewing after fixes, stay within your dimension: focus ONLY on
|
|
428
|
+
the previously flagged scope plus regressions introduced by the fixes;
|
|
429
|
+
confirm each prior finding item by item (fixed / not fixed / partial).
|
|
361
430
|
`,
|
|
362
431
|
color: "#FB923C", // orange
|
|
363
432
|
// May edit ONLY the blackboard artifact; never the reviewed code
|
|
@@ -398,6 +467,17 @@ pass/fail signal.
|
|
|
398
467
|
- **ENVIRONMENT** — tooling/deps/config issue. Report precisely; do not
|
|
399
468
|
work around silently.
|
|
400
469
|
|
|
470
|
+
## UI verification mode (when the dispatch targets user-visible behavior)
|
|
471
|
+
- Drive the real page/flow when tooling allows (browser automation, dev
|
|
472
|
+
server): verify core user flows with real data — create, view,
|
|
473
|
+
interact, navigate — and capture screenshots plus console output as
|
|
474
|
+
evidence. Terminal-only API checks are not UI verification.
|
|
475
|
+
- If no browser tooling is available in this environment, do not fake
|
|
476
|
+
it: run whatever is verifiable (build, static checks, API-level
|
|
477
|
+
tests) and end your report with the line
|
|
478
|
+
\`UI NOT VERIFIED: <what still needs manual checking>\`
|
|
479
|
+
so the lead can relay it to the user.
|
|
480
|
+
|
|
401
481
|
## Output format
|
|
402
482
|
- Test files created/modified.
|
|
403
483
|
- Run command used and result: passed / failed / error counts.
|
|
@@ -457,13 +537,34 @@ by the team — flag prominently if that is the case.
|
|
|
457
537
|
- Separate fact from interpretation explicitly.
|
|
458
538
|
- Prefer official documentation; quote the relevant lines when reading code.
|
|
459
539
|
- State which product/version each finding applies to.
|
|
540
|
+
|
|
541
|
+
## Behavioral constraints
|
|
542
|
+
- When analyzing dependencies, output call-graph diagrams in mermaid format.
|
|
543
|
+
- When the code under study involves authentication/authorization, tag each
|
|
544
|
+
finding with a security-risk level (Critical / High / Medium / Low).
|
|
545
|
+
- When encountering unfamiliar modules, note what additional context would
|
|
546
|
+
help and suggest which tool or command could retrieve it.
|
|
547
|
+
- Before starting deep analysis, check whether a design doc or prior
|
|
548
|
+
research artifact already exists in the project — reference it instead
|
|
549
|
+
of re-deriving.
|
|
460
550
|
`,
|
|
461
551
|
color: "#A78BFA", // violet
|
|
462
552
|
permission: { edit: "allow", bash: "allow", webfetch: "allow", websearch: "allow" },
|
|
463
553
|
};
|
|
554
|
+
/* Shared rules appended to every specialist prompt (after the blackboard). */
|
|
555
|
+
const SHARED_RULES = `
|
|
556
|
+
|
|
557
|
+
## Evidence rule
|
|
558
|
+
Every "done / fixed / passed" claim in your reply must carry its
|
|
559
|
+
evidence: command output, log lines, a diff, a screenshot. No
|
|
560
|
+
narrative-only completions.
|
|
561
|
+
|
|
562
|
+
## Project conventions
|
|
563
|
+
If the project README (or AGENTS.md) is in your Reads, treat its
|
|
564
|
+
conventions as binding — they outrank your defaults.`;
|
|
464
565
|
/* Append the blackboard rules to every specialist prompt. */
|
|
465
566
|
for (const a of [architect, implementer, reviewer, tester, researcher]) {
|
|
466
|
-
a.prompt = (a.prompt ?? "") + BLACKBOARD_GUARANTEE;
|
|
567
|
+
a.prompt = (a.prompt ?? "") + BLACKBOARD_GUARANTEE + SHARED_RULES;
|
|
467
568
|
}
|
|
468
569
|
/* ------------------------------------------------------------------ */
|
|
469
570
|
/* Export all agents keyed by name */
|
package/dist/agents.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"agents.js","sourceRoot":"","sources":["../src/agents.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAIH,wEAAwE;AACxE,+DAA+D;AAC/D,wEAAwE;AACxE,MAAM,oBAAoB,GAAG;;;;;;;;;;;;;;;;;;;;oCAoBO,CAAA;AAEpC,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,QAAQ,GAAgB;IAC5B,IAAI,EAAE,SAAS;IACf,WAAW,EACT,2EAA2E;QAC3E,qEAAqE;QACrE,2EAA2E;QAC3E,2EAA2E;QAC3E,4BAA4B;IAC9B,MAAM,EAAE
|
|
1
|
+
{"version":3,"file":"agents.js","sourceRoot":"","sources":["../src/agents.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAIH,wEAAwE;AACxE,+DAA+D;AAC/D,wEAAwE;AACxE,MAAM,oBAAoB,GAAG;;;;;;;;;;;;;;;;;;;;oCAoBO,CAAA;AAEpC,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,QAAQ,GAAgB;IAC5B,IAAI,EAAE,SAAS;IACf,WAAW,EACT,2EAA2E;QAC3E,qEAAqE;QACrE,2EAA2E;QAC3E,2EAA2E;QAC3E,4BAA4B;IAC9B,MAAM,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAsNT;IACC,KAAK,EAAE,SAAS,EAAE,SAAS;IAC3B,UAAU,EAAE;QACV,IAAI,EAAE,OAAO;QACb,IAAI,EAAE,OAAO;QACb,QAAQ,EAAE,OAAO;QACjB,IAAI,EAAE,OAAO;KACd;CACF,CAAA;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,SAAS,GAAgB;IAC7B,IAAI,EAAE,UAAU;IAChB,WAAW,EACT,2EAA2E;QAC3E,2EAA2E;QAC3E,0EAA0E;QAC1E,yCAAyC;IAC3C,MAAM,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAmCT;IACC,KAAK,EAAE,SAAS,EAAE,WAAW;IAC7B,sEAAsE;IACtE,mCAAmC;IACnC,UAAU,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE;CAC/D,CAAA;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,WAAW,GAAgB;IAC/B,IAAI,EAAE,UAAU;IAChB,WAAW,EACT,uEAAuE;QACvE,0EAA0E;QAC1E,gEAAgE;IAClE,MAAM,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAiCT;IACC,KAAK,EAAE,SAAS,EAAE,QAAQ;IAC1B,UAAU,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE;CAChE,CAAA;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,QAAQ,GAAgB;IAC5B,IAAI,EAAE,UAAU;IAChB,WAAW,EACT,qEAAqE;QACrE,uEAAuE;QACvE,wEAAwE;QACxE,4CAA4C;IAC9C,MAAM,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAkDT;IACC,KAAK,EAAE,SAAS,EAAE,SAAS;IAC3B,iEAAiE;IACjE,uCAAuC;IACvC,UAAU,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE;CAChE,CAAA;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,MAAM,GAAgB;IAC1B,IAAI,EAAE,UAAU;IAChB,WAAW,EACT,qEAAqE;QACrE,yEAAyE;QACzE,0DAA0D;IAC5D,MAAM,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAiDT;IACC,KAAK,EAAE,SAAS,EAAE,OAAO;IACzB,UAAU,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE;CAChE,CAAA;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AACxE,MAAM,UAAU,GAAgB;IAC9B,IAAI,EAAE,UAAU;IAChB,WAAW,EACT,iEAAiE;QACjE,sEAAsE;QACtE,yEAAyE;QACzE,mCAAmC;IACrC,MAAM,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CA4CT;IACC,KAAK,EAAE,SAAS,EAAE,SAAS;IAC3B,UAAU,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE;CACpF,CAAA;AAED,8EAA8E;AAC9E,MAAM,YAAY,GAAG;;;;;;;;;qDASgC,CAAA;AAErD,6DAA6D;AAC7D,KAAK,MAAM,CAAC,IAAI,CAAC,SAAS,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,EAAE,UAAU,CAAC,EAAE,CAAC;IACvE,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,GAAG,oBAAoB,GAAG,YAAY,CAAA;AACnE,CAAC;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AAExE,MAAM,CAAC,MAAM,MAAM,GAAgC;IACjD,MAAM,EAAE,QAAQ;IAChB,SAAS;IACT,WAAW;IACX,QAAQ;IACR,MAAM;IACN,UAAU;CACX,CAAA"}
|
package/dist/commands.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"commands.d.ts","sourceRoot":"","sources":["../src/commands.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;
|
|
1
|
+
{"version":3,"file":"commands.d.ts","sourceRoot":"","sources":["../src/commands.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AA+J/C,eAAO,MAAM,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,aAAa,CAOlD,CAAA"}
|
package/dist/commands.js
CHANGED
|
@@ -40,16 +40,22 @@ $ARGUMENTS
|
|
|
40
40
|
};
|
|
41
41
|
/* ------------------------------------------------------------------ */
|
|
42
42
|
const teamReview = {
|
|
43
|
-
description: "Review code
|
|
43
|
+
description: "Review code with a single focused dimension — completeness, correctness, or impact (default: correctness).",
|
|
44
44
|
agent: "reviewer",
|
|
45
|
-
template: `Perform a
|
|
45
|
+
template: `Perform a single-dimension code review on the following scope.
|
|
46
46
|
|
|
47
47
|
## Scope
|
|
48
48
|
$ARGUMENTS
|
|
49
49
|
|
|
50
|
+
## Dimension
|
|
51
|
+
$ARGUMENTS may name one dimension: completeness (requirements coverage),
|
|
52
|
+
correctness (logic & security), or impact (regressions & blast radius).
|
|
53
|
+
If no dimension is named, review correctness. Ignore the other dimensions
|
|
54
|
+
— parallel reviewers own them.
|
|
55
|
+
|
|
50
56
|
If no specific scope is given, review all recently modified files in the project.
|
|
51
57
|
|
|
52
|
-
Use the standard
|
|
58
|
+
Use the standard severity scale:
|
|
53
59
|
- 🔴 Critical (must fix)
|
|
54
60
|
- 🟡 Warning (should fix)
|
|
55
61
|
- 🔵 Suggestion (nice to have)
|
|
@@ -104,7 +110,9 @@ $ARGUMENTS
|
|
|
104
110
|
1. Triage first: if this is a question/consult rather than an action
|
|
105
111
|
request, answer it without touching files (propose fixes, wait for a
|
|
106
112
|
go-ahead). Honor and restate any user-stated boundaries everywhere.
|
|
107
|
-
2.
|
|
113
|
+
2. Read the project's README first (plus AGENTS.md/CLAUDE.md if present) —
|
|
114
|
+
extract binding conventions and restate them in every dispatch. Then
|
|
115
|
+
understand the goal and acceptance criteria.
|
|
108
116
|
3. **Create a todo list FIRST** — one item per work package, with checkable
|
|
109
117
|
done-conditions. Update statuses live (exactly one in_progress).
|
|
110
118
|
4. Dispatch sub-tasks to the appropriate agents:
|
|
@@ -113,6 +121,10 @@ $ARGUMENTS
|
|
|
113
121
|
- Code review → reviewer
|
|
114
122
|
- Testing → tester
|
|
115
123
|
- Research → researcher
|
|
124
|
+
Respect the pipeline gates: research completes before planning, design
|
|
125
|
+
before code, code before verify, verify before review, UI verification
|
|
126
|
+
before "done" on user-visible frontend changes. Batch independent
|
|
127
|
+
dispatches into the same round.
|
|
116
128
|
5. Coordinate through the shared blackboard: this conversation owns one
|
|
117
129
|
session folder (<root>/<session-key>/ — compact timestamp, created on
|
|
118
130
|
first board write), one task directory per task inside it; every
|
|
@@ -120,14 +132,19 @@ $ARGUMENTS
|
|
|
120
132
|
Write to: one owned artifact file); artifacts are frozen — revisions are
|
|
121
133
|
new round-suffixed files; keep MANIFEST.md's \`## Current state\` header
|
|
122
134
|
updated each converged round. Run independent sub-tasks in parallel.
|
|
123
|
-
6. Enforce
|
|
124
|
-
|
|
125
|
-
|
|
126
|
-
|
|
127
|
-
|
|
128
|
-
|
|
129
|
-
|
|
130
|
-
|
|
135
|
+
6. Enforce Ultra Review for non-trivial changes: exactly 3 parallel
|
|
136
|
+
reviewer dispatches, one dimension each (completeness / correctness /
|
|
137
|
+
impact), then merge the three reports into one severity-grouped list.
|
|
138
|
+
7. Enforce the feedback loop: merged Critical/Major findings and tester
|
|
139
|
+
product-bugs become fix tasks on the list → implementer fixes →
|
|
140
|
+
re-review affected scope → re-run tests. Repeat until clean (max 2
|
|
141
|
+
loops, then escalate).
|
|
142
|
+
8. Collect all outputs, resolve conflicts, synthesize a final result.
|
|
143
|
+
9. Present a structured summary: changes, review/test verdict, remaining
|
|
144
|
+
assumptions and risks. If the project keeps a CHANGELOG.md, append an
|
|
145
|
+
entry for the delivered changes (offer to create one if missing). Leave
|
|
146
|
+
the task blackboard directory in place — the plugin's TTL sweeper
|
|
147
|
+
reclaims idle boards; never delete it yourself.`,
|
|
131
148
|
};
|
|
132
149
|
/* ------------------------------------------------------------------ */
|
|
133
150
|
/* Export all commands keyed by name */
|
package/dist/commands.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"commands.js","sourceRoot":"","sources":["../src/commands.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,wEAAwE;AACxE,MAAM,QAAQ,GAAkB;IAC9B,WAAW,EACT,+FAA+F;IACjG,KAAK,EAAE,WAAW;IAClB,QAAQ,EAAE;;;;;;;;;;;;;kDAasC;CACjD,CAAA;AAED,wEAAwE;AACxE,MAAM,aAAa,GAAkB;IACnC,WAAW,EACT,0FAA0F;IAC5F,KAAK,EAAE,aAAa;IACpB,QAAQ,EAAE;;;;;;;;4DAQgD;CAC3D,CAAA;AAED,wEAAwE;AACxE,MAAM,UAAU,GAAkB;IAChC,WAAW,EACT,
|
|
1
|
+
{"version":3,"file":"commands.js","sourceRoot":"","sources":["../src/commands.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,wEAAwE;AACxE,MAAM,QAAQ,GAAkB;IAC9B,WAAW,EACT,+FAA+F;IACjG,KAAK,EAAE,WAAW;IAClB,QAAQ,EAAE;;;;;;;;;;;;;kDAasC;CACjD,CAAA;AAED,wEAAwE;AACxE,MAAM,aAAa,GAAkB;IACnC,WAAW,EACT,0FAA0F;IAC5F,KAAK,EAAE,aAAa;IACpB,QAAQ,EAAE;;;;;;;;4DAQgD;CAC3D,CAAA;AAED,wEAAwE;AACxE,MAAM,UAAU,GAAkB;IAChC,WAAW,EACT,4GAA4G;IAC9G,KAAK,EAAE,UAAU;IACjB,QAAQ,EAAE;;;;;;;;;;;;;;;;;;;gEAmBoD;CAC/D,CAAA;AAED,wEAAwE;AACxE,MAAM,QAAQ,GAAkB;IAC9B,WAAW,EACT,2EAA2E;IAC7E,KAAK,EAAE,QAAQ;IACf,QAAQ,EAAE;;;;;;;;;;;8BAWkB;CAC7B,CAAA;AAED,wEAAwE;AACxE,MAAM,YAAY,GAAkB;IAClC,WAAW,EACT,uEAAuE;IACzE,KAAK,EAAE,YAAY;IACnB,QAAQ,EAAE;;;;;;;;;;;0DAW8C;CACzD,CAAA;AAED,wEAAwE;AACxE,MAAM,OAAO,GAAkB;IAC7B,WAAW,EACT,mGAAmG;IACrG,KAAK,EAAE,MAAM;IACb,QAAQ,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;mDA2CuC;CAClD,CAAA;AAED,wEAAwE;AACxE,wEAAwE;AACxE,wEAAwE;AAExE,MAAM,CAAC,MAAM,QAAQ,GAAkC;IACrD,WAAW,EAAE,QAAQ;IACrB,gBAAgB,EAAE,aAAa;IAC/B,aAAa,EAAE,UAAU;IACzB,WAAW,EAAE,QAAQ;IACrB,eAAe,EAAE,YAAY;IAC7B,UAAU,EAAE,OAAO;CACpB,CAAA"}
|
package/dist/index.d.ts
CHANGED
|
@@ -17,13 +17,12 @@
|
|
|
17
17
|
* Options (via the tuple plugin form):
|
|
18
18
|
* "plugin": [["@te-river/opencode-team-mode@latest", { "ttlDays": 7 }]]
|
|
19
19
|
* - `ttlDays` → blackboard auto-cleanup TTL; default 5.
|
|
20
|
-
* - `defaultAgent` →
|
|
21
|
-
*
|
|
22
|
-
*
|
|
23
|
-
*
|
|
24
|
-
*
|
|
25
|
-
*
|
|
26
|
-
* are mutually exclusive).
|
|
20
|
+
* - `defaultAgent` → Team is the default agent (opt-out: set `false`).
|
|
21
|
+
* Owning the default slot means the picker pins it
|
|
22
|
+
* FIRST — order becomes team, build, plan. Set
|
|
23
|
+
* `false` for build-as-default with Team in its
|
|
24
|
+
* alphabetical slot: build, plan, team (the two are
|
|
25
|
+
* mutually exclusive by the server's sort).
|
|
27
26
|
*/
|
|
28
27
|
import type { OpenCodePlugin } from "./types.js";
|
|
29
28
|
declare const plugin: OpenCodePlugin;
|
package/dist/index.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAEH,OAAO,KAAK,EAAE,cAAc,EAAkB,MAAM,YAAY,CAAA;AA2BhE,QAAA,MAAM,MAAM,EAAE,cAyCb,CAAA;AAED,eAAe,MAAM,CAAA"}
|
package/dist/index.js
CHANGED
|
@@ -17,13 +17,12 @@
|
|
|
17
17
|
* Options (via the tuple plugin form):
|
|
18
18
|
* "plugin": [["@te-river/opencode-team-mode@latest", { "ttlDays": 7 }]]
|
|
19
19
|
* - `ttlDays` → blackboard auto-cleanup TTL; default 5.
|
|
20
|
-
* - `defaultAgent` →
|
|
21
|
-
*
|
|
22
|
-
*
|
|
23
|
-
*
|
|
24
|
-
*
|
|
25
|
-
*
|
|
26
|
-
* are mutually exclusive).
|
|
20
|
+
* - `defaultAgent` → Team is the default agent (opt-out: set `false`).
|
|
21
|
+
* Owning the default slot means the picker pins it
|
|
22
|
+
* FIRST — order becomes team, build, plan. Set
|
|
23
|
+
* `false` for build-as-default with Team in its
|
|
24
|
+
* alphabetical slot: build, plan, team (the two are
|
|
25
|
+
* mutually exclusive by the server's sort).
|
|
27
26
|
*/
|
|
28
27
|
import { agents } from "./agents.js";
|
|
29
28
|
import { commands } from "./commands.js";
|
|
@@ -77,8 +76,8 @@ const plugin = {
|
|
|
77
76
|
continue;
|
|
78
77
|
cfg.command[name] = def;
|
|
79
78
|
}
|
|
80
|
-
// ---------- make Team the default agent (opt-
|
|
81
|
-
const promote = options?.defaultAgent
|
|
79
|
+
// ---------- make Team the default agent (opt-out: defaultAgent:false)
|
|
80
|
+
const promote = options?.defaultAgent !== false;
|
|
82
81
|
if (promote && (!cfg.default_agent || cfg.default_agent === "build")) {
|
|
83
82
|
cfg.default_agent = "team";
|
|
84
83
|
}
|
package/dist/index.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;GAyBG;AAGH,OAAO,EAAE,MAAM,EAAE,MAAM,aAAa,CAAA;AACpC,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAA;AACxC,OAAO,EACL,0BAA0B,EAC1B,YAAY,EACZ,gBAAgB,GACjB,MAAM,iBAAiB,CAAA;AAExB,iEAAiE;AACjE,SAAS,cAAc,CAAC,IAAY,EAAE,OAAe;IACnD,OAAO;QACL,EAAE;QACF,EAAE;QACF,kDAAkD;QAClD,qBAAqB,IAAI,IAAI;QAC7B,qFAAqF;QACrF,qFAAqF;QACrF,uFAAuF;QACvF,qFAAqF;QACrF,yDAAyD;QACzD,sGAAsG;QACtG,yEAAyE,OAAO,gCAAgC;QAChH,sFAAsF;KACvF,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACd,CAAC;AAED,MAAM,MAAM,GAAmB;IAC7B,EAAE,EAAE,WAAW;IAEf,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE;QAC/B,wEAAwE;QACxE,MAAM,SAAS,GACb,OAAO,KAAK,EAAE,SAAS,KAAK,QAAQ,IAAI,KAAK,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC;YAChE,CAAC,CAAC,KAAK,CAAC,SAAS;YACjB,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,CAAA;QACnB,MAAM,KAAK,GAAG,YAAY,CAAC,OAAO,CAAC,CAAA;QACnC,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,CAAC,IAAI,gBAAgB,CAAA;QAC7E,MAAM,SAAS,GAAG,0BAA0B,CAAC,SAAS,EAAE,KAAK,CAAC,CAAA;QAC9D,MAAM,IAAI,GAAG,cAAc,CAAC,SAAS,EAAE,OAAO,CAAC,CAAA;QAE/C,OAAO;YACL,iEAAiE;YACjE,MAAM,CAAC,GAAmB;gBACxB,IAAI,CAAC,GAAG,CAAC,KAAK;oBAAE,GAAG,CAAC,KAAK,GAAG,EAAE,CAAA;gBAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;oBACjD,mEAAmE;oBACnE,IAAI,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC;wBAAE,SAAQ;oBAC7B,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG;wBAChB,GAAG,GAAG;wBACN,MAAM,EAAE,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM;qBACjE,CAAA;gBACH,CAAC;gBAED,IAAI,CAAC,GAAG,CAAC,OAAO;oBAAE,GAAG,CAAC,OAAO,GAAG,EAAE,CAAA;gBAClC,KAAK,MAAM,CAAC,IAAI,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;oBACnD,IAAI,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC;wBAAE,SAAQ;oBAC/B,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,GAAG,CAAA;gBACzB,CAAC;gBAED,uEAAuE;gBACvE,MAAM,OAAO,GAAG,OAAO,EAAE,YAAY,KAAK,KAAK,CAAA;gBAC/C,IAAI,OAAO,IAAI,CAAC,CAAC,GAAG,CAAC,aAAa,IAAI,GAAG,CAAC,aAAa,KAAK,OAAO,CAAC,EAAE,CAAC;oBACrE,GAAG,CAAC,aAAa,GAAG,MAAM,CAAA;gBAC5B,CAAC;YACH,CAAC;SACF,CAAA;IACH,CAAC;CACF,CAAA;AAED,eAAe,MAAM,CAAA"}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@te-river/opencode-team-mode",
|
|
3
|
-
"version": "1.4.
|
|
3
|
+
"version": "1.4.6",
|
|
4
4
|
"description": "Team collaboration mode plugin for OpenCode Desktop — injects specialized agents (architect, reviewer, implementer, tester, researcher) and team workflow commands into your OpenCode workspace.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"main": "dist/index.js",
|