You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: .github/workflows/tests.yml
+1-1Lines changed: 1 addition & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -37,5 +37,5 @@ jobs:
37
37
run: >
38
38
uv run pytest massgen/tests
39
39
-m "not live_api and not docker and not expensive"
40
-
-k "not test_timeline_snapshot and not test_final_lock_option and not test_web_quickstart_reasoning_sync and not test_subagent_input_bar_snapshot_matches_main_input"
40
+
-k "not test_timeline_snapshot and not test_final_lock_option and not test_web_quickstart_reasoning_sync and not test_subagent_input_bar_snapshot_matches_main_input and not test_review_modal_snapshot"
Trace analyzer subagents now launch in the background after each round to write insights from execution traces into memory. Improved evaluation criteria generation and system prompt tuning. Fixes for final injection, eval criteria GPT pre-collab, trace analyzer launch, and trace memory.
WebUI automation now auto-starts without browser interaction — open the URL at any point mid-run to monitor progress. MassGen skill redesign for increased usability and WebUI integration. Quickstart Wizard rework, Workspace Browser expansion, and flexible evaluation criteria field names.
17
20
18
-
**v0.1.68 (March 25, 2026)** - Checkpoint Mode
19
-
New checkpoint coordination mode with delegator pattern — main agent plans solo then delegates to team via `checkpoint()` tool. LLM API circuit breaker for 429 handling. WebUI checkpoint support. LiteLLM supply chain fix.
21
+
---
22
+
23
+
## [0.1.71] - 2026-04-01
24
+
25
+
### Changed
26
+
-**Better Evaluation Criteria**: Improved criteria generation for higher-quality, more opinionated output
27
+
-**System Prompt Tuning**: Adjusted system prompts for better agent performance across coordination rounds
28
+
29
+
### Fixed
30
+
-**Final Injection Fix**: Corrected injection behavior at the final stage
31
+
-**Eval Criteria GPT Pre-Collab Fix**: Resolved evaluation criteria issues with GPT models during pre-collaboration phase
32
+
-**Execution Trace Analyzer Launch Fix**: Trace analyzer now starts correctly
33
+
-**Trace Memory Fix**: Corrected memory handling in execution traces
34
+
-**Auto Round Memory Fix**: Fixed automatic round handling for memory
Copy file name to clipboardExpand all lines: CONTRIBUTING.md
+4-4Lines changed: 4 additions & 4 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -359,7 +359,7 @@ Create a `.env` file in the `massgen` directory as described in [README](README.
359
359
360
360
## 🔧 Development Workflow
361
361
362
-
> **Important**: Our next version is v0.1.71. If you want to contribute, please contribute to the `dev/v0.1.71` branch (or `main` if dev/v0.1.71 doesn't exist yet).
362
+
> **Important**: Our next version is v0.1.72. If you want to contribute, please contribute to the `dev/v0.1.72` branch (or `main` if dev/v0.1.72 doesn't exist yet).
363
363
364
364
### 1. Create Feature Branch
365
365
@@ -368,7 +368,7 @@ Create a `.env` file in the `massgen` directory as described in [README](README.
368
368
git fetch upstream
369
369
370
370
# Create feature branch from dev/v0.1.60 (or main if dev branch doesn't exist yet)
@@ -155,21 +155,20 @@ This project started with the "threads of thought" and "iterative refinement" id
155
155
156
156
---
157
157
158
-
## 🆕 Latest Features (v0.1.70)
158
+
## 🆕 Latest Features (v0.1.71)
159
159
160
-
**🎉 Released: March 30, 2026**
160
+
**🎉 Released: April 1, 2026**
161
161
162
-
**What's New in v0.1.70:**
163
-
-**📋 Evaluation Criteria Redesign** - Three-tier categorization (`primary`, `standard`, `stretch`) with anti-pattern definitions and aspiration statements.
164
-
-**🔄 Improved Checklist-Gated Evaluation** - Tighter iterative submission cycles with improved scoring and improvement proposals.
165
-
-**⚡ Fast Iteration Mode** - Streamlined multi-round submission phases via `fast_iteration.yaml`.
166
-
-**🔍 WebUI Review Modal** - Approve and comment on outputs directly in the browser.
162
+
**What's New in v0.1.71:**
163
+
-**🔍 Trace Analyzer Subagents** - Launch in the background after each round to write insights from execution traces into memory.
164
+
-**📋 Better Evaluation Criteria** - Improved criteria generation for higher-quality, more opinionated output.
165
+
-**🧠 System Prompt Tuning** - Adjusted system prompts for better agent performance across coordination rounds.
166
+
-**🔧 Stability Fixes** - Fixed final injection, eval criteria GPT pre-collab, trace analyzer launch, and memory handling.
167
167
168
-
**Try v0.1.70 Features:**
168
+
**Try v0.1.71 Features:**
169
169
```bash
170
-
pip install massgen==0.1.70
171
-
# Try fast iteration with redesigned evaluation criteria
172
-
uv run massgen --config @examples/features/fast_iteration.yaml "Create an svg of an AI agent coding."
170
+
pip install massgen==0.1.71
171
+
uv run massgen --config @examples/features/trace_analyzer_background.yaml "Create an svg of an AI agent coding."
173
172
```
174
173
175
174
→ [See full release history and examples](massgen/configs/README.md#release-history--examples)
@@ -1241,18 +1240,19 @@ MassGen is currently in its foundational stage, with a focus on parallel, asynch
1241
1240
1242
1241
⚠️ **Early Stage Notice:** As MassGen is in active development, please expect upcoming breaking architecture changes as we continue to refine and improve the system.
1243
1242
1244
-
### Recent Achievements (v0.1.70)
1243
+
### Recent Achievements (v0.1.71)
1245
1244
1246
-
**🎉 Released: March 30, 2026**
1245
+
**🎉 Released: April 1, 2026**
1247
1246
1248
-
#### Evaluation Criteria Redesign
1249
-
-**Evaluation Criteria Redesign** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Three-tier categorization (`primary`, `standard`, `stretch`) with anti-pattern definitions and aspiration statements
1250
-
-**Improved Checklist-Gated Evaluation** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Tighter iterative submission cycles with improved scoring and improvement proposals
1251
-
-**Fast Iteration Mode** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Streamlined multi-round submission phases via `fast_iteration.yaml`
1252
-
-**WebUI Review Modal** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Approve and comment on outputs directly in the browser when working in git
-**Trace Analyzer Subagents**: Background trace analysis after each round — writes insights from execution traces into memory for next-round continuity
1249
+
-**Better Evaluation Criteria**: Improved criteria generation for higher-quality, more opinionated output
1250
+
-**System Prompt Tuning**: Adjusted system prompts for better agent performance across coordination rounds
1251
+
-**Stability Fixes**: Fixed final injection, eval criteria GPT pre-collab, trace analyzer launch, trace memory, and auto round memory
1254
1252
1255
-
### Previous Achievements (v0.0.3 - v0.1.69)
1253
+
### Previous Achievements (v0.0.3 - v0.1.70)
1254
+
1255
+
✅ **Evaluation Criteria Redesign (v0.1.70)**: Redesigned three-tier evaluation criteria with anti-pattern definitions and aspiration statements. Improved checklist-gated evaluation. Fast iteration mode, WebUI review modal, and background trace analysis.
1256
1256
1257
1257
✅ **WebUI Automation & Improved Skill (v0.1.69)**: WebUI automation auto-starts without browser interaction. MassGen skill redesign for increased usability and WebUI integration. Quickstart Wizard rework and Workspace Browser expansion.
1258
1258
@@ -1537,9 +1537,9 @@ MassGen is currently in its foundational stage, with a focus on parallel, asynch
1537
1537
1538
1538
We welcome community contributions to achieve these goals.
1539
1539
1540
-
### v0.1.71 Roadmap
1540
+
### v0.1.72 Roadmap
1541
1541
1542
-
Version 0.1.71 focuses on cloud execution:
1542
+
Version 0.1.72 focuses on cloud execution:
1543
1543
1544
1544
#### Planned Features
1545
1545
-**Cloud Modal MVP** ([#982](https://github.com/massgen/MassGen/issues/982)): Run MassGen as a cloud job on Modal — progress streams to terminal, results saved locally under `.massgen/cloud_jobs/`
@@ -154,21 +154,20 @@ This project started with the "threads of thought" and "iterative refinement" id
154
154
155
155
---
156
156
157
-
## 🆕 Latest Features (v0.1.70)
157
+
## 🆕 Latest Features (v0.1.71)
158
158
159
-
**🎉 Released: March 30, 2026**
159
+
**🎉 Released: April 1, 2026**
160
160
161
-
**What's New in v0.1.70:**
162
-
-**📋 Evaluation Criteria Redesign** - Three-tier categorization (`primary`, `standard`, `stretch`) with anti-pattern definitions and aspiration statements.
163
-
-**🔄 Improved Checklist-Gated Evaluation** - Tighter iterative submission cycles with improved scoring and improvement proposals.
164
-
-**⚡ Fast Iteration Mode** - Streamlined multi-round submission phases via `fast_iteration.yaml`.
165
-
-**🔍 WebUI Review Modal** - Approve and comment on outputs directly in the browser.
161
+
**What's New in v0.1.71:**
162
+
-**🔍 Trace Analyzer Subagents** - Launch in the background after each round to write insights from execution traces into memory.
163
+
-**📋 Better Evaluation Criteria** - Improved criteria generation for higher-quality, more opinionated output.
164
+
-**🧠 System Prompt Tuning** - Adjusted system prompts for better agent performance across coordination rounds.
165
+
-**🔧 Stability Fixes** - Fixed final injection, eval criteria GPT pre-collab, trace analyzer launch, and memory handling.
166
166
167
-
**Try v0.1.70 Features:**
167
+
**Try v0.1.71 Features:**
168
168
```bash
169
-
pip install massgen==0.1.70
170
-
# Try fast iteration with redesigned evaluation criteria
171
-
uv run massgen --config @examples/features/fast_iteration.yaml "Create an svg of an AI agent coding."
169
+
pip install massgen==0.1.71
170
+
uv run massgen --config @examples/features/trace_analyzer_background.yaml "Create an svg of an AI agent coding."
172
171
```
173
172
174
173
→ [See full release history and examples](massgen/configs/README.md#release-history--examples)
@@ -1240,18 +1239,19 @@ MassGen is currently in its foundational stage, with a focus on parallel, asynch
1240
1239
1241
1240
⚠️ **Early Stage Notice:** As MassGen is in active development, please expect upcoming breaking architecture changes as we continue to refine and improve the system.
1242
1241
1243
-
### Recent Achievements (v0.1.70)
1242
+
### Recent Achievements (v0.1.71)
1244
1243
1245
-
**🎉 Released: March 30, 2026**
1244
+
**🎉 Released: April 1, 2026**
1246
1245
1247
-
#### Evaluation Criteria Redesign
1248
-
-**Evaluation Criteria Redesign** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Three-tier categorization (`primary`, `standard`, `stretch`) with anti-pattern definitions and aspiration statements
1249
-
-**Improved Checklist-Gated Evaluation** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Tighter iterative submission cycles with improved scoring and improvement proposals
1250
-
-**Fast Iteration Mode** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Streamlined multi-round submission phases via `fast_iteration.yaml`
1251
-
-**WebUI Review Modal** ([#1035](https://github.com/massgen/MassGen/pull/1035)): Approve and comment on outputs directly in the browser when working in git
-**Trace Analyzer Subagents**: Background trace analysis after each round — writes insights from execution traces into memory for next-round continuity
1248
+
-**Better Evaluation Criteria**: Improved criteria generation for higher-quality, more opinionated output
1249
+
-**System Prompt Tuning**: Adjusted system prompts for better agent performance across coordination rounds
1250
+
-**Stability Fixes**: Fixed final injection, eval criteria GPT pre-collab, trace analyzer launch, trace memory, and auto round memory
1253
1251
1254
-
### Previous Achievements (v0.0.3 - v0.1.69)
1252
+
### Previous Achievements (v0.0.3 - v0.1.70)
1253
+
1254
+
✅ **Evaluation Criteria Redesign (v0.1.70)**: Redesigned three-tier evaluation criteria with anti-pattern definitions and aspiration statements. Improved checklist-gated evaluation. Fast iteration mode, WebUI review modal, and background trace analysis.
1255
1255
1256
1256
✅ **WebUI Automation & Improved Skill (v0.1.69)**: WebUI automation auto-starts without browser interaction. MassGen skill redesign for increased usability and WebUI integration. Quickstart Wizard rework and Workspace Browser expansion.
1257
1257
@@ -1536,9 +1536,9 @@ MassGen is currently in its foundational stage, with a focus on parallel, asynch
1536
1536
1537
1537
We welcome community contributions to achieve these goals.
1538
1538
1539
-
### v0.1.71 Roadmap
1539
+
### v0.1.72 Roadmap
1540
1540
1541
-
Version 0.1.71 focuses on cloud execution:
1541
+
Version 0.1.72 focuses on cloud execution:
1542
1542
1543
1543
#### Planned Features
1544
1544
-**Cloud Modal MVP** ([#982](https://github.com/massgen/MassGen/issues/982)): Run MassGen as a cloud job on Modal — progress streams to terminal, results saved locally under `.massgen/cloud_jobs/`
This roadmap outlines MassGen's development priorities for upcoming releases. Each release focuses on specific capabilities with real-world use cases.
10
10
@@ -42,14 +42,26 @@ Want to contribute or collaborate on a specific track? Reach out to the track ow
42
42
43
43
| Release | Target | Feature | Owner | Use Case |
44
44
|---------|--------|---------|-------|----------|
45
-
|**v0.1.71**| 04/02/26 | Cloud Modal MVP |@ncrispino| Run MassGen as a cloud job on Modal ([#982](https://github.com/massgen/MassGen/issues/982)) |
46
-
|**v0.1.72**| 04/04/26 | OpenAI Audio API |@ncrispino| Support OpenAI audio API for audio understanding ([#960](https://github.com/massgen/MassGen/issues/960)) |
47
-
|**v0.1.73**| 04/07/26 | Image/Video Edit Capabilities |@ncrispino| Check and support img/video editing capabilities ([#959](https://github.com/massgen/MassGen/issues/959)) |
45
+
|**v0.1.72**| 04/04/26 | Cloud Modal MVP |@ncrispino| Run MassGen as a cloud job on Modal ([#982](https://github.com/massgen/MassGen/issues/982)) |
46
+
|**v0.1.73**| 04/07/26 | OpenAI Audio API |@ncrispino| Support OpenAI audio API for audio understanding ([#960](https://github.com/massgen/MassGen/issues/960)) |
47
+
|**v0.1.74**| 04/09/26 | Image/Video Edit Capabilities |@ncrispino| Check and support img/video editing capabilities ([#959](https://github.com/massgen/MassGen/issues/959)) |
-**Trace Analyzer Subagents**: Background trace analysis after each round — writes insights from execution traces into memory for next-round continuity
59
+
-**Better Evaluation Criteria**: Improved criteria generation for higher-quality, more opinionated output
60
+
-**System Prompt Tuning**: Adjusted system prompts for better agent performance across coordination rounds
61
+
-**Stability Fixes**: Fixed final injection, eval criteria GPT pre-collab, trace analyzer launch, trace memory, and auto round memory
0 commit comments