HelixTranslate/docs/Issues.html at main · HelixDevelopment/HelixTranslate · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
  <meta charset="utf-8" />
  <meta name="generator" content="pandoc" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes" />
  <title>Issues</title>
  <style>
    /* Default styles provided by pandoc.
    ** See https://pandoc.org/MANUAL.html#variables-for-html for config info.
    */
    html {
      color: #1a1a1a;
      background-color: #fdfdfd;
    }
    body {
      margin: 0 auto;
      max-width: 36em;
      padding-left: 50px;
      padding-right: 50px;
      padding-top: 50px;
      padding-bottom: 50px;
      hyphens: auto;
      overflow-wrap: break-word;
      text-rendering: optimizeLegibility;
      font-kerning: normal;
    }
    @media (max-width: 600px) {
      body {
        font-size: 0.9em;
        padding: 12px;
      }
      h1 {
        font-size: 1.8em;
      }
    }
    @media print {
      html {
        background-color: white;
      }
      body {
        background-color: transparent;
        color: black;
        font-size: 12pt;
      }
      p, h2, h3 {
        orphans: 3;
        widows: 3;
      }
      h2, h3, h4 {
        page-break-after: avoid;
      }
    }
    p {
      margin: 1em 0;
    }
    a {
      color: #1a1a1a;
    }
    a:visited {
      color: #1a1a1a;
    }
    img {
      max-width: 100%;
    }
    svg {
      height: auto;
      max-width: 100%;
    }
    h1, h2, h3, h4, h5, h6 {
      margin-top: 1.4em;
    }
    h5, h6 {
      font-size: 1em;
      font-style: italic;
    }
    h6 {
      font-weight: normal;
    }
    ol, ul {
      padding-left: 1.7em;
      margin-top: 1em;
    }
    li > ol, li > ul {
      margin-top: 0;
    }
    blockquote {
      margin: 1em 0 1em 1.7em;
      padding-left: 1em;
      border-left: 2px solid #e6e6e6;
      color: #606060;
    }
    code {
      white-space: pre-wrap;
      font-family: Menlo, Monaco, Consolas, 'Lucida Console', monospace;
      font-size: 85%;
      margin: 0;
      hyphens: manual;
    }
    pre {
      margin: 1em 0;
      overflow: auto;
    }
    pre code {
      padding: 0;
      overflow: visible;
      overflow-wrap: normal;
    }
    .sourceCode {
     background-color: transparent;
     overflow: visible;
    }
    hr {
      border: none;
      border-top: 1px solid #1a1a1a;
      height: 1px;
      margin: 1em 0;
    }
    table {
      margin: 1em 0;
      border-collapse: collapse;
      width: 100%;
      overflow-x: auto;
      display: block;
      font-variant-numeric: lining-nums tabular-nums;
    }
    table caption {
      margin-bottom: 0.75em;
    }
    tbody {
      margin-top: 0.5em;
      border-top: 1px solid #1a1a1a;
      border-bottom: 1px solid #1a1a1a;
    }
    th {
      border-top: 1px solid #1a1a1a;
      padding: 0.25em 0.5em 0.25em 0.5em;
    }
    td {
      padding: 0.125em 0.5em 0.25em 0.5em;
    }
    header {
      margin-bottom: 4em;
      text-align: center;
    }
    #TOC li {
      list-style: none;
    }
    #TOC ul {
      padding-left: 1.3em;
    }
    #TOC > ul {
      padding-left: 0;
    }
    #TOC a:not(:hover) {
      text-decoration: none;
    }
    span.smallcaps{font-variant: small-caps;}
    div.columns{display: flex; gap: min(4vw, 1.5em);}
    div.column{flex: auto; overflow-x: auto;}
    div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
    /* The extra [class] is a hack that increases specificity enough to
       override a similar rule in reveal.js */
    ul.task-list[class]{list-style: none;}
    ul.task-list li input[type="checkbox"] {
      font-size: inherit;
      width: 0.8em;
      margin: 0 0.8em 0.2em -1.6em;
      vertical-align: middle;
    }
    .display.math{display: block; text-align: center; margin: 0.5rem auto;}
  </style>
</head>
<body>
<header id="title-block-header">
<h1 class="title">Issues</h1>
</header>
<h1 id="helixtranslate--issues-open-workable-items">HelixTranslate —
Issues (Open Workable Items)</h1>
<p><strong>Revision:</strong> 1 <strong>Last modified:</strong>
2026-06-14T15:37:53Z <strong>Authority:</strong> §11.4.15 (status) ·
§11.4.16 (type) · §11.4.19 (column-alignment) · §11.4.21
(Operator-blocked details) · §11.4.54 (ATM-NNN ticket IDs) · §11.4.91
(clear descriptions) <strong>Scope:</strong> the open-work tracker.
Every entry carries a stable <code>[ATM-NNN]</code> id, a
<code>**Status:**</code>, a <code>**Type:**</code>, and — for
Operator-blocked items — an <code>**Operator-Block-Details:**</code>
line (§11.4.21). Source of truth for the open items is
<code>docs/WORKING_PLAN.md</code>; this file is the
constitution-mandated tracker view of it.</p>
<p>Status vocabulary (open): <code>Queued</code> ·
<code>Operator-blocked</code> · <code>Design</code> (architecture
decision pending) · <code>Blocked</code> (another session / external
dependency).</p>
<p>ATM ids continue the monotonic sequence after
<code>docs/Fixed.md</code> (last allocated ATM-064).</p>
<hr />
<h3
id="1-atm-065-decide-the-single-authoritative-version-number-23x-vs-300">§1.
[ATM-065] Decide the single authoritative version number (2.3.x vs
3.0.0)</h3>
<p><strong>Status:</strong> Operator-blocked <strong>Type:</strong>
Task</p>
<ul>
<li>WHAT: <code>VERSION</code>=2.3.0 but Makefile/binaries historically
referenced 3.0.0. The single-source wiring landed (ATM-064); the actual
number to ship is an operator decision.</li>
<li><strong>Operator-Block-Details:</strong> WHAT — pick 2.3.x or 3.0.0
as the next release version. WHY — CLAUDE.md says treat
<code>VERSION</code> (2.3.0) as authoritative while the Makefile
references 3.0.0; the agent cannot decide product versioning.
Self-resolution exhausted: single-source wiring already done (a36030e),
value choice is purely a product call. UNBLOCK CONDITION — operator
states the canonical number; agent sets <code>VERSION</code> and re-runs
the version test. WHO — operator (Milos Vasic).</li>
</ul>
<h3
id="2-atm-066-provider-credentials-absentinvalid-openaianthropicgeminizhipu-block-allowlist-refresh">§2.
[ATM-066] Provider credentials absent/invalid
(OpenAI/Anthropic/Gemini/Zhipu) block allowlist refresh</h3>
<p><strong>Status:</strong> Operator-blocked <strong>Type:</strong>
Bug</p>
<ul>
<li>WHAT: OPENAI_API_KEY and ANTHROPIC_API_KEY absent → those providers
unverified and their model allowlists stale; GEMINI_API_KEY invalid
(live <code>/models</code> returns "API Key not found"); Zhipu account
out of balance (error 1113 余额不足) so its current models
(glm-4.5/4.6/5/5.1…) cannot be translation-verified.</li>
<li><strong>Operator-Block-Details:</strong> WHAT — add/refresh
OPENAI/ANTHROPIC/GEMINI keys and recharge the Zhipu account. WHY — the
proven deepseek pattern (live <code>/models</code> → verify-translate →
additive allowlist + RED gate) requires a working key to verify the
response shape; refreshing an allowlist without verification would be a
§11.4 bluff. Self-resolution exhausted: live <code>/models</code> probes
captured, deepseek fix <code>0fd1a34</code> is the ready template.
UNBLOCK CONDITION — funded/valid keys present in the environment. WHO —
operator (Milos Vasic).</li>
</ul>
<h3
id="3-atm-067-30-other-provider-allowlists-not-audited-against-live-current-models">§3.
[ATM-067] ~30 other provider allowlists not audited against live current
models</h3>
<p><strong>Status:</strong> Operator-blocked <strong>Type:</strong>
Task</p>
<ul>
<li>WHAT: qwen, groq, cohere, mistral, xai, replicate, cerebras,
cloudflare, siliconflow, hyperbolic, togetherai, sambanova, kimi,
novita, nlpcloud, upstage, sarvam, modal, publicai, nia, vulavula —
allowlists unverified against current live catalogs.</li>
<li><strong>Operator-Block-Details:</strong> WHAT — provide funded keys
per provider. WHY — each allowlist refresh needs a live verify-translate
+ string-content shape check before adding (deepseek pattern); without
keys the verification cannot run. Self-resolution exhausted: pattern
proven, only credentials missing. UNBLOCK CONDITION — per-provider
funded keys. WHO — operator (Milos Vasic).</li>
</ul>
<h3
id="4-atm-068-inert-cli-flags-in-unified-translator--chunk-size-workers-concurrency-verify--monitoring-stub">§4.
[ATM-068] Inert CLI flags in unified-translator
(-chunk-size/-workers/-concurrency/-verify, -monitoring stub)</h3>
<p><strong>Status:</strong> Design <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: <code>cmd/unified-translator</code> parses
<code>-chunk-size</code>, <code>-workers</code>,
<code>-concurrency</code>, <code>-verify</code> but never consumes them
(chunking is automatic+correct via
<code>translateWithRetry</code>/<code>splitText</code>);
<code>startMonitoringServer</code> is a print-only stub. Decision
needed: wire each flag to real semantics OR remove it (removing a
user-facing flag needs operator confirmation per §11.4.122). Not a blind
autonomous fix.</li>
</ul>
<h3
id="5-atm-069-inert-config-fields-docxconfigmintextlengthignorestyles-pdfconfigmintextlength">§5.
[ATM-069] Inert config fields (DOCXConfig.MinTextLength/IgnoreStyles,
PDFConfig.MinTextLength)</h3>
<p><strong>Status:</strong> Design <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: <code>DOCXConfig.MinTextLength</code>,
<code>DOCXConfig.IgnoreStyles</code>, and
<code>PDFConfig.MinTextLength</code> are declared/documented but never
consumed. Decision: wire (filter short paragraphs / honor ignore-styles)
or drop. Wiring <code>MinTextLength</code> changes output, so it needs
care + tests.</li>
</ul>
<h3
id="6-atm-070-verifier-minscorethreshold-scale-inconsistency-0-100-raw-vs-0-10-normalized">§6.
[ATM-070] Verifier MinScoreThreshold scale inconsistency (0-100 raw vs
0-10 normalized)</h3>
<p><strong>Status:</strong> Design <strong>Type:</strong> Bug</p>
<ul>
<li>WHAT: the handler (<code>pkg/api/verifier_handlers.go</code>)
compares <code>MinScoreThreshold</code> against raw 0-100
<code>OverallScore</code>; the adapter
(<code>internal/services/llmsverifier_score_adapter.go</code>) compares
it against the normalized 0-10 score. The two contracts contradict and
<code>GetPreferences</code>/<code>GetProviderScore</code> have no
production caller. Canonical scale must be declared before either path
is wired (fixing one side blindly breaks the other's test,
§11.4.120).</li>
</ul>
<h3
id="7-atm-071-reasoning-model-structured-content-support-content-as-list-not-string">§7.
[ATM-071] Reasoning-model structured-content support (content as LIST,
not STRING)</h3>
<p><strong>Status:</strong> Design <strong>Type:</strong> Feature</p>
<ul>
<li>WHAT: OpenAI-compatible clients assume <code>content</code> is a
string; some reasoning models return <code>content</code> as a
structured list (verified Mistral <code>magistral-medium-latest</code>;
likely glm-5 / deepseek-reasoner class), which the clients silently
drop. Adding structured-content handling would unlock those models.
Non-trivial; design + tests required.</li>
</ul>
<h3 id="8-atm-072-markdown-not-a-first-class-cli-input-format">§8.
[ATM-072] Markdown not a first-class CLI input format</h3>
<p><strong>Status:</strong> Design <strong>Type:</strong> Feature</p>
<ul>
<li>WHAT: <code>.md</code> input is detected as TXT and translated as
plain text (works, but markdown structure is not preserved). First-class
markdown input that preserves structure is an enhancement.</li>
</ul>
<h3
id="9-atm-073-cmdtranslator-intermediate-markdown-download-dir-inconsistency-needs-live-ssh">§9.
[ATM-073] cmd/translator intermediate-markdown download-dir
inconsistency (needs live SSH)</h3>
<p><strong>Status:</strong> Blocked <strong>Type:</strong> Bug</p>
<ul>
<li>WHAT: intermediate <code>.md</code> downloads to
<code>Dir(InputFile)</code> in one path vs <code>Dir(OutputFile)</code>
in another; manifests only under live SSH with <code>-o</code> in a
different dir. Not unit-testable without real SSH infra; reproduce via
the §11.4.76 Containers submodule (boot an SSH worker container) → RED →
fix → GREEN. Gated on containerized SSH test infrastructure.</li>
</ul>
<h3
id="10-atm-074-pkghash-is-a-dead-package-zero-importers--investigate-per-114124">§10.
[ATM-074] pkg/hash is a dead package (zero importers) — investigate per
§11.4.124</h3>
<p><strong>Status:</strong> Operator-blocked <strong>Type:</strong>
Task</p>
<ul>
<li>WHAT: <code>pkg/hash</code> (393 LOC) is a <code>package main</code>
duplicate of <code>pkg/version.CodebaseHasher</code> with zero importers
(confirmed; documented in commit <code>981ced9</code>). Per §11.4.124 it
must not be removed without git-history investigation; per §11.4.122
removing a shipped package needs operator confirmation.</li>
<li><strong>Operator-Block-Details:</strong> WHAT — confirm
keep-or-remove of the dead <code>pkg/hash</code> package. WHY —
§11.4.122 forbids silently removing an existing component, and §11.4.124
requires operator confirmation before deleting a shipped package even
with git-history proof. Self-resolution exhausted: investigation done
(981ced9 captures it as a dead duplicate). UNBLOCK CONDITION — operator
says remove (then a separate descriptive commit cites the evidence) or
keep/wire-in. WHO — operator (Milos Vasic).</li>
</ul>
<h3 id="11-atm-075-pre-build-cm--gate-suite-not-implemented">§11.
[ATM-075] Pre-build CM-* gate suite not implemented</h3>
<p><strong>Status:</strong> Queued <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: the constitution references dozens of <code>CM-*</code>
pre-build gates + paired §1.1 mutations; this project has only
<code>scripts/testing/meta_test_constitution_inheritance.sh</code> and
partial seeds. Implement the highest-value gates first (anti-bluff
smoke, doc-sync, regression-guard-registered, no-fakes-beyond-unit,
gitignore-precommit-audit), each with a paired mutation, wired into
<code>scripts/pre_build_verification.sh</code>. (P4.2 — a sibling
session owns <code>pre_build_verification.sh</code>.)</li>
</ul>
<h3
id="12-atm-076-11465-universal-markdown-export-audit-across-all-tracked-docs">§12.
[ATM-076] §11.4.65 universal markdown export audit across all tracked
docs</h3>
<p><strong>Status:</strong> Queued <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: every tracked non-source <code>.md</code> must have synced
<code>.html</code>+<code>.pdf</code> siblings. The commit wrapper
auto-syncs, but a full audit (including WORKING_PLAN.md and all docs/)
is needed to confirm every <code>docs/*.md</code> has fresh
exports.</li>
</ul>
<h3
id="13-atm-077-owned-submodule-bug-hunt--brittle-test-fixes-11428-equal-codebase">§13.
[ATM-077] Owned-submodule bug-hunt + brittle-test fixes (§11.4.28
equal-codebase)</h3>
<p><strong>Status:</strong> Blocked <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: owned submodules (challenges, containers, helix_qa,
doc_processor, llm_orchestrator, llm_provider, vision_engine,
llms_verifier, docs_chain) are §11.4.28 equal-codebase and carry the
brittle "connection refused" env-coupled test class fixed in the main
module this session.</li>
<li>WHY-BLOCKED: evidence shows another session is actively working
<code>helix_qa</code> (<code>go test -race</code> observed); per
§11.4.119 single-owner do not collide. After operator confirms
ownership/coordination, run per-submodule bug-hunt waves (separate
go.mod, separate upstreams).</li>
</ul>
<h3
id="14-atm-078-per-feature-test-type-matrix--helixqa--challenges-coverage-1142511427">§14.
[ATM-078] Per-feature test-type matrix + HelixQA + Challenges coverage
(§11.4.25/§11.4.27)</h3>
<p><strong>Status:</strong> Queued <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: §11.4.27 mandates 100% coverage with every test type. This
session added unit/integration + some stress/chaos + real E2E proofs;
the full matrix per feature (perf/benchmark for the pipeline, chaos for
distributed/storage, Challenges entries per shipped feature, full
HelixQA autonomous sessions) is unfilled. Build the §11.4.25 coverage
ledger and fill highest-value gaps.</li>
</ul>
<h3 id="15-atm-079-docsqa-evidence-per-shipped-feature-11483">§15.
[ATM-079] docs/qa/<run-id> evidence per shipped feature (§11.4.83)</h3>
<p><strong>Status:</strong> Queued <strong>Type:</strong> Task</p>
<ul>
<li>WHAT: §11.4.83 requires a recorded e2e transcript per shipped
feature. E2E proofs exist for PDF input, the output-format matrix, and
deepseek-v4; the remaining user-visible features from this session's
fixes need per-feature evidence dirs under <code>docs/qa/</code>.</li>
</ul>
<h3 id="16-atm-080-full-11440-7-step-release-retest-not-yet-run">§16.
[ATM-080] Full §11.4.40 7-step release retest not yet run</h3>
<p><strong>Status:</strong> Operator-blocked <strong>Type:</strong>
Task</p>
<ul>
<li>WHAT: release requires the complete §11.4.40 7-step retest
(pre-build sweep, post-build sweep, on-device cycle, meta-test mutation
sweep, Challenge bank sweep, Issues/Fixed audit, CONTINUATION sync) on a
clean baseline. Repeated full <code>go test ./... -p 1</code> sweeps are
green but the full 7-step ritual is not done.</li>
<li><strong>Operator-Block-Details:</strong> WHAT — authorize and run
the full release retest. WHY — sequenced after the gate suite (ATM-075),
Challenges/HelixQA (ATM-078), and §11.4.151 release-prefix tagging;
requires operator confirmation of the release scope. Self-resolution
exhausted: dependencies tracked as their own items. UNBLOCK CONDITION —
ATM-075/ATM-078 closed + operator green-lights the release. WHO —
operator (Milos Vasic).</li>
</ul>
<h3 id="17-atm-081-no-114151-prefixed-release-tag-yet">§17. [ATM-081] No
§11.4.151 prefixed release tag yet</h3>
<p><strong>Status:</strong> Operator-blocked <strong>Type:</strong>
Task</p>
<ul>
<li>WHAT: no release tag exists. Per §11.4.151 the tag must be
<code>helix_translate-&lt;version&gt;</code> (prefix from
<code>HELIX_RELEASE_PREFIX</code> env or lowercased root dir name) on
the main repo + every owned submodule, fanned to all upstreams with no
force-push (§11.4.113).</li>
<li><strong>Operator-Block-Details:</strong> WHAT — green-light creating
the prefixed release tag. WHY — gated on ATM-080 (the full retest must
pass first) and the operator-chosen version (ATM-065); a tag cannot
precede a clean retest per §11.4.40. Self-resolution exhausted: tag
mechanics ready, only the go-ahead + passing retest missing. UNBLOCK
CONDITION — ATM-080 GREEN + ATM-065 decided. WHO — operator (Milos
Vasic).</li>
</ul>
<hr />
<p><em>This tracker is regenerated into <code>Issues_Summary.md</code>
by <code>scripts/testing/generate_issues_summary.sh</code>. Do not
hand-edit the summary. When an item closes it migrates atomically to
<code>Fixed.md</code> (§11.4.19).</em></p>
</body>
</html>