Skip to content

Commit 9b2af5e

Browse files
lhy0718lhy0718
committed
Remove advanced details section from multiple language README files, streamlining content to focus on project overview and documentation references.
1 parent 66307f0 commit 9b2af5e

10 files changed

Lines changed: 18 additions & 1216 deletions

File tree

‎README.md‎

Lines changed: 8 additions & 131 deletions
Original file line numberDiff line numberDiff line change
@@ -31,7 +31,7 @@
3131
<a href="./docs/README.ru.md"><strong>Русский</strong></a>
3232
</p>
3333

34-
<p><sub>Localized README files are maintained translations of this document. For normative wording and latest edits, use the English README as the canonical reference.</sub></p>
34+
<p><sub>Localized README files are maintained translations of this document. The English README is updated first.</sub></p>
3535

3636
<p>
3737
<a href="https://github.com/lhy0718/AutoLabOS/actions/workflows/ci.yml">
@@ -99,7 +99,7 @@ In practice:
9999
4. Weak evidence triggers backtracking or downgrade instead of automatic polishing.
100100
5. If the review gate passes, `write_paper` drafts a manuscript from bounded evidence.
101101

102-
The historical 9-node contract remains the architectural baseline. In the current runtime, `figure_audit` is the one approved post-analysis checkpoint inserted between `analyze_results` and `review` so figure-quality critique can checkpoint and resume independently.
102+
In the current runtime, `figure_audit` sits between `analyze_results` and `review` so figure-quality critique can checkpoint and resume independently.
103103

104104
```mermaid
105105
stateDiagram-v2
@@ -196,7 +196,7 @@ The brief is not just a startup note. It is the governed contract for a run.
196196

197197
That makes the brief part of the audit trail, not just part of the prompt.
198198

199-
In the current contract, `.autolabos/config.yaml` is primarily for provider/runtime defaults and workspace policy. Run-specific research intent, evidence bars, baseline expectations, manuscript-format targets, and manuscript template path belong in the brief. Persisted config may therefore omit brief-owned sections such as research defaults and some manuscript-profile or paper-template fields.
199+
In practice, `.autolabos/config.yaml` holds provider and workspace defaults, while the brief carries run-specific research intent, evidence bars, baseline expectations, manuscript-format targets, and manuscript template path.
200200

201201
```bash
202202
/new
@@ -288,12 +288,14 @@ Failure fingerprints are persisted so structural errors and repeated equivalent
288288

289289
### Reproducibility Through Artifacts
290290

291+
Runs stay inspectable because the system persists artifacts, checkpoints, and transitions instead of relying on hidden state.
292+
291293

292294
---
293295

294-
## Validation And Harness-Oriented Quality Model
296+
## Quality Model
295297

296-
AutoLabOS treats validation surfaces as first-class.
298+
AutoLabOS makes quality checks visible during a run.
297299

298300
- `/doctor` checks environment and workspace readiness before a run starts
299301

@@ -415,131 +417,6 @@ AutoLabOS also has built-in harness presets such as `base`, `compact`, `failure-
415417

416418
---
417419

418-
## Advanced Details
419-
420-
<details>
421-
<summary><strong>Execution modes</strong></summary>
422-
423-
AutoLabOS preserves the governed workflow and safety gates across every mode.
424-
425-
| Mode | Command | Behavior |
426-
|---|---|---|
427-
| **Interactive** | `autolabos` | Slash-command TUI with explicit approval gates |
428-
| **Minimal approval** | Config: `approval_mode: minimal` | Auto-approves safe transitions |
429-
| **Hybrid approval** | Config: `approval_mode: hybrid` | Auto-advances strong low-risk transitions, pauses risky or low-confidence ones |
430-
| **Overnight** | `/agent overnight [run]` | Unattended single pass, 24-hour limit, conservative backtracking |
431-
| **Autonomous** | `/agent autonomous [run]` | Open-ended bounded research exploration |
432-
433-
</details>
434-
435-
<details>
436-
<summary><strong>Governance artifact flow</strong></summary>
437-
438-
```mermaid
439-
flowchart LR
440-
Brief["Research Brief<br/>completeness artifact"] --> Design["design_experiments"]
441-
Design --> Contract["Experiment Contract<br/>hypothesis, single change,<br/>confound check"]
442-
Design --> Consistency["Brief-Design Consistency<br/>warnings artifact"]
443-
Contract --> Run["run_experiments"]
444-
Run --> Failures["Failure Memory<br/>fingerprinted JSONL"]
445-
Run --> Analyze["analyze_results"]
446-
Analyze --> Decision["Attempt Decision<br/>keep/discard/replicate"]
447-
Decision --> FigureAudit["figure_audit"]
448-
FigureAudit --> Review["review"]
449-
Failures --> Review
450-
Contract --> Review
451-
Review --> Ceiling["Pre-Review Summary<br/>claim ceiling detail"]
452-
Ceiling --> Paper["write_paper"]
453-
```
454-
455-
</details>
456-
457-
<details>
458-
<summary><strong>Artifact flow</strong></summary>
459-
460-
```mermaid
461-
flowchart TB
462-
A["collect_papers"] --> A1["corpus.jsonl, bibtex.bib"]
463-
A1 --> B["analyze_papers"]
464-
B --> B1["paper_summaries.jsonl, evidence_store.jsonl"]
465-
B1 --> C["generate_hypotheses"]
466-
C --> C1["hypotheses.jsonl"]
467-
C1 --> D["design_experiments"]
468-
D --> D1["experiment_plan.yaml, experiment_contract.json,<br/>brief_design_consistency.json"]
469-
D1 --> E["implement_experiments"]
470-
E --> F["run_experiments"]
471-
F --> F1["metrics.json, failure_memory.jsonl,<br/>objective_evaluation.json"]
472-
F1 --> G["analyze_results"]
473-
G --> G1["result_analysis.json, attempt_decisions.jsonl,<br/>transition_recommendation.json"]
474-
G1 --> H["figure_audit"]
475-
H --> H1["gate1_gate2_issues.json,<br/>figure_audit_summary.json"]
476-
H1 --> I["review"]
477-
I --> I1["pre_review_summary.json, review_packet.json,<br/>minimum_gate.json, paper_critique.json"]
478-
I1 --> J["write_paper"]
479-
J --> J1["main.tex, references.bib,<br/>scientific_validation.json, main.pdf"]
480-
```
481-
482-
</details>
483-
484-
<details>
485-
<summary><strong>Node architecture</strong></summary>
486-
487-
| Node | Role(s) | What it does |
488-
|---|---|---|
489-
| `collect_papers` | collector, curator | Discovers and curates candidate paper set via Semantic Scholar |
490-
| `analyze_papers` | reader, evidence extractor | Extracts summaries and evidence from selected papers |
491-
| `generate_hypotheses` | hypothesis agent + skeptical reviewer | Synthesizes ideas from literature, then pressure-tests them |
492-
| `design_experiments` | designer + feasibility/statistical/ops panel | Filters plans for practicality, writes experiment contract |
493-
| `implement_experiments` | implementer | Produces code and workspace changes through ACI actions |
494-
| `run_experiments` | runner + failure triager + rerun planner | Drives execution, records failures, decides reruns |
495-
| `analyze_results` | analyst + metric auditor + confounder detector | Checks result reliability, writes attempt decisions |
496-
| `figure_audit` | figure auditor + optional vision critique | Checks evidence alignment, captions/references, and publication readiness before review |
497-
| `review` | 5-specialist panel + claim ceiling + two-layer gate | Structural review - blocks writing if evidence is insufficient |
498-
| `write_paper` | paper writer + reviewer critique | Drafts manuscript, runs post-draft critique, builds PDF |
499-
500-
</details>
501-
502-
<details>
503-
<summary><strong>Bounded automation</strong></summary>
504-
505-
| Node | Internal automation | Bound |
506-
|---|---|---|
507-
| `analyze_papers` | Auto-expands evidence window when too sparse | <= 2 expansions |
508-
| `design_experiments` | Deterministic panel scoring + experiment contract | Runs once per design |
509-
| `run_experiments` | Failure triage + one-shot transient rerun | Never retries structural failures |
510-
| `run_experiments` | Failure memory fingerprinting | >= 3 identical exhausts retries |
511-
| `analyze_results` | Objective rematching + result panel calibration | One rematch before human pause |
512-
| `figure_audit` | Gate 3 figure critique + summary aggregation | Vision critique remains independently resumable |
513-
| `write_paper` | Related-work scout + validation-aware repair | 1 repair pass max |
514-
515-
</details>
516-
517-
<details>
518-
<summary><strong>Public output bundle</strong></summary>
519-
520-
```
521-
outputs/<title-slug>-<run_id_prefix>/
522-
├── paper/
523-
├── experiment/
524-
├── analysis/
525-
├── review/
526-
├── results/
527-
├── reproduce/
528-
├── manifest.json
529-
└── README.md
530-
```
531-
532-
</details>
533-
534-
---
535-
536420
## Status
537421

538-
AutoLabOS is an active OSS research-engineering project. The canonical references for behavior and contracts are the repository docs under `docs/`, especially:
539-
540-
- `docs/architecture.md`
541-
- `docs/experiment-quality-bar.md`
542-
- `docs/paper-quality-bar.md`
543-
- `docs/reproducibility.md`
544-
- `docs/research-brief-template.md`
545-
422+
AutoLabOS is an active OSS research-engineering project. For deeper details beyond this overview, see the documents under docs.

‎docs/README.de.md‎

Lines changed: 2 additions & 78 deletions
Original file line numberDiff line numberDiff line change
@@ -417,82 +417,6 @@ AutoLabOS bietet außerdem built-in harness presets wie `base`, `compact`, `fail
417417

418418
---
419419

420-
## Advanced Details
420+
## Status
421421

422-
<details>
423-
<summary><strong>Ausführungsmodi</strong></summary>
424-
425-
AutoLabOS behält den governed workflow und die safety gates in allen Modi bei.
426-
427-
| Modus | Befehl | Verhalten |
428-
|---|---|---|
429-
| **Interactive** | `autolabos` | Slash-Command-TUI mit expliziten Approval Gates |
430-
| **Minimal approval** | Config: `approval_mode: minimal` | Genehmigt sichere Übergänge automatisch |
431-
| **Hybrid approval** | Config: `approval_mode: hybrid` | Lässt starke und risikoarme Übergänge automatisch weiterlaufen; pausiert riskante oder wenig vertrauenswürdige Übergänge |
432-
| **Overnight** | `/agent overnight [run]` | Unbeaufsichtigte Ein-Pass-Ausführung, 24-Stunden-Limit, konservatives Backtracking |
433-
| **Autonomous** | `/agent autonomous [run]` | Open-ended bounded research exploration |
434-
435-
</details>
436-
437-
<details>
438-
<summary><strong>Governance Artifact Flow</strong></summary>
439-
440-
```mermaid
441-
flowchart LR
442-
Brief["Research Brief<br/>completeness artifact"] --> Design["design_experiments"]
443-
Design --> Contract["Experiment Contract<br/>hypothesis, single change,<br/>confound check"]
444-
Design --> Consistency["Brief-Design Consistency<br/>warnings artifact"]
445-
Contract --> Run["run_experiments"]
446-
Run --> Failures["Failure Memory<br/>fingerprinted JSONL"]
447-
Run --> Analyze["analyze_results"]
448-
Analyze --> Decision["Attempt Decision<br/>keep/discard/replicate"]
449-
Decision --> FigureAudit["figure_audit"]
450-
FigureAudit --> Review["review"]
451-
Failures --> Review
452-
Contract --> Review
453-
Review --> Ceiling["Pre-Review Summary<br/>claim ceiling detail"]
454-
Ceiling --> Paper["write_paper"]
455-
```
456-
457-
</details>
458-
459-
<details>
460-
<summary><strong>Artifact Flow</strong></summary>
461-
462-
```mermaid
463-
flowchart TB
464-
A["collect_papers"] --> A1["corpus.jsonl, bibtex.bib"]
465-
A1 --> B["analyze_papers"]
466-
B --> B1["paper_summaries.jsonl, evidence_store.jsonl"]
467-
B1 --> C["generate_hypotheses"]
468-
C --> C1["hypotheses.jsonl"]
469-
C1 --> D["design_experiments"]
470-
D --> D1["experiment_plan.yaml, experiment_contract.json,<br/>brief_design_consistency.json"]
471-
D1 --> E["implement_experiments"]
472-
E --> F["run_experiments"]
473-
F --> F1["metrics.json, failure_memory.jsonl,<br/>objective_evaluation.json"]
474-
F1 --> G["analyze_results"]
475-
G --> G1["result_analysis.json, attempt_decisions.jsonl,<br/>transition_recommendation.json"]
476-
G1 --> H["figure_audit"]
477-
H --> H1["gate1_gate2_issues.json,<br/>figure_audit_summary.json"]
478-
H1 --> I["review"]
479-
I --> I1["pre_review_summary.json, review_packet.json,<br/>minimum_gate.json, paper_critique.json"]
480-
I1 --> J["write_paper"]
481-
J --> J1["main.tex, references.bib,<br/>scientific_validation.json, main.pdf"]
482-
```
483-
484-
</details>
485-
486-
<details>
487-
<summary><strong>Status</strong></summary>
488-
489-
AutoLabOS ist ein aktiv entwickeltes OSS-Research-Engineering-Projekt. Die kanonischen Referenzen für Verhalten und Contracts liegen unter `docs/`, insbesondere:
490-
491-
- `docs/architecture.md`
492-
- `docs/experiment-quality-bar.md`
493-
- `docs/paper-quality-bar.md`
494-
- `docs/reproducibility.md`
495-
- `docs/research-brief-template.md`
496-
497-
498-
</details>
422+
AutoLabOS ist ein aktives OSS-Projekt für Research Engineering. Weitere Details finden sich in den Dokumenten unter `docs/`.

‎docs/README.es.md‎

Lines changed: 1 addition & 126 deletions
Original file line numberDiff line numberDiff line change
@@ -417,131 +417,6 @@ AutoLabOS también tiene built-in harness presets como `base`, `compact`, `failu
417417

418418
---
419419

420-
## Advanced Details
421-
422-
<details>
423-
<summary><strong>Modos de ejecución</strong></summary>
424-
425-
AutoLabOS mantiene el governed workflow y los safety gates en todos los modos.
426-
427-
| Modo | Comando | Comportamiento |
428-
|---|---|---|
429-
| **Interactive** | `autolabos` | TUI con slash commands y approval gates explícitos |
430-
| **Minimal approval** | Config: `approval_mode: minimal` | Autoaprueba transiciones seguras |
431-
| **Hybrid approval** | Config: `approval_mode: hybrid` | Avanza automáticamente en transiciones fuertes y de bajo riesgo; pausa las de riesgo o baja confianza |
432-
| **Overnight** | `/agent overnight [run]` | Ejecución unattended de una sola pasada, límite de 24h, backtracking conservador |
433-
| **Autonomous** | `/agent autonomous [run]` | Open-ended bounded research exploration |
434-
435-
</details>
436-
437-
<details>
438-
<summary><strong>Governance artifact flow</strong></summary>
439-
440-
```mermaid
441-
flowchart LR
442-
Brief["Research Brief<br/>completeness artifact"] --> Design["design_experiments"]
443-
Design --> Contract["Experiment Contract<br/>hypothesis, single change,<br/>confound check"]
444-
Design --> Consistency["Brief-Design Consistency<br/>warnings artifact"]
445-
Contract --> Run["run_experiments"]
446-
Run --> Failures["Failure Memory<br/>fingerprinted JSONL"]
447-
Run --> Analyze["analyze_results"]
448-
Analyze --> Decision["Attempt Decision<br/>keep/discard/replicate"]
449-
Decision --> FigureAudit["figure_audit"]
450-
FigureAudit --> Review["review"]
451-
Failures --> Review
452-
Contract --> Review
453-
Review --> Ceiling["Pre-Review Summary<br/>claim ceiling detail"]
454-
Ceiling --> Paper["write_paper"]
455-
```
456-
457-
</details>
458-
459-
<details>
460-
<summary><strong>Artifact flow</strong></summary>
461-
462-
```mermaid
463-
flowchart TB
464-
A["collect_papers"] --> A1["corpus.jsonl, bibtex.bib"]
465-
A1 --> B["analyze_papers"]
466-
B --> B1["paper_summaries.jsonl, evidence_store.jsonl"]
467-
B1 --> C["generate_hypotheses"]
468-
C --> C1["hypotheses.jsonl"]
469-
C1 --> D["design_experiments"]
470-
D --> D1["experiment_plan.yaml, experiment_contract.json,<br/>brief_design_consistency.json"]
471-
D1 --> E["implement_experiments"]
472-
E --> F["run_experiments"]
473-
F --> F1["metrics.json, failure_memory.jsonl,<br/>objective_evaluation.json"]
474-
F1 --> G["analyze_results"]
475-
G --> G1["result_analysis.json, attempt_decisions.jsonl,<br/>transition_recommendation.json"]
476-
G1 --> H["figure_audit"]
477-
H --> H1["gate1_gate2_issues.json,<br/>figure_audit_summary.json"]
478-
H1 --> I["review"]
479-
I --> I1["pre_review_summary.json, review_packet.json,<br/>minimum_gate.json, paper_critique.json"]
480-
I1 --> J["write_paper"]
481-
J --> J1["main.tex, references.bib,<br/>scientific_validation.json, main.pdf"]
482-
```
483-
484-
</details>
485-
486-
<details>
487-
<summary><strong>Arquitectura de nodos</strong></summary>
488-
489-
| Nodo | Rol | Qué hace |
490-
|---|---|---|
491-
| `collect_papers` | collector, curator | Descubre y filtra candidate paper sets con Semantic Scholar |
492-
| `analyze_papers` | reader, evidence extractor | Extrae summaries y evidence de los papers seleccionados |
493-
| `generate_hypotheses` | hypothesis agent + skeptical reviewer | Sintetiza ideas desde la literatura y las somete a presión |
494-
| `design_experiments` | designer + feasibility/statistical/ops panel | Filtra planes por viabilidad y escribe el experiment contract |
495-
| `implement_experiments` | implementer | Produce cambios de código y workspace mediante ACI actions |
496-
| `run_experiments` | runner + failure triager + rerun planner | Ejecuta experimentos, registra fallos y decide reruns |
497-
| `analyze_results` | analyst + metric auditor + confounder detector | Comprueba la fiabilidad de resultados y escribe attempt decisions |
498-
| `figure_audit` | figure auditor + optional vision critique | Comprueba evidence alignment, captions / references y publication readiness |
499-
| `review` | 5-specialist panel + claim ceiling + two-layer gate | Realiza structural review y bloquea la escritura si falta evidencia |
500-
| `write_paper` | paper writer + reviewer critique | Redacta el manuscrito, ejecuta post-draft critique y construye el PDF |
501-
502-
</details>
503-
504-
<details>
505-
<summary><strong>Bounded automation</strong></summary>
506-
507-
| Nodo | Automatización interna | Límite |
508-
|---|---|---|
509-
| `analyze_papers` | Autoexpansión de evidence window cuando la evidencia es escasa | <= 2 expansiones |
510-
| `design_experiments` | Deterministic panel scoring + experiment contract | Una vez por diseño |
511-
| `run_experiments` | Failure triage + un rerun transitorio | Nunca reintenta fallos estructurales |
512-
| `run_experiments` | Failure memory fingerprinting | >= 3 fallos idénticos agotan los retries |
513-
| `analyze_results` | Objective rematching + result panel calibration | Un rematch antes de pausa humana |
514-
| `figure_audit` | Gate 3 figure critique + summary aggregation | La vision critique sigue siendo resumable de forma independiente |
515-
| `write_paper` | Related-work scout + validation-aware repair | 1 repair como máximo |
516-
517-
</details>
518-
519-
<details>
520-
<summary><strong>Public output bundle</strong></summary>
521-
522-
```
523-
outputs/<title-slug>-<run_id_prefix>/
524-
├── paper/
525-
├── experiment/
526-
├── analysis/
527-
├── review/
528-
├── results/
529-
├── reproduce/
530-
├── manifest.json
531-
└── README.md
532-
```
533-
534-
</details>
535-
536-
---
537-
538420
## Estado
539421

540-
AutoLabOS es un proyecto OSS activo de research engineering. Las referencias canónicas de comportamiento y contracts están en `docs/`, especialmente:
541-
542-
- `docs/architecture.md`
543-
- `docs/experiment-quality-bar.md`
544-
- `docs/paper-quality-bar.md`
545-
- `docs/reproducibility.md`
546-
- `docs/research-brief-template.md`
547-
422+
AutoLabOS es un proyecto OSS activo de research engineering. Si necesitas más detalle, consulta los documentos del directorio docs.

0 commit comments

Comments
 (0)