@fro.bot/systematic 1.12.0 → 1.13.0

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.
Files changed (11) hide show
  1. package/README.md +3 -3
  2. package/commands/workflows/brainstorm.md +17 -8
  3. package/commands/workflows/compound.md +95 -60
  4. package/commands/workflows/plan.md +22 -24
  5. package/commands/workflows/review.md +45 -32
  6. package/commands/workflows/work.md +91 -19
  7. package/dist/cli.js +1 -1
  8. package/dist/{index-0ftaxvrt.js → index-bky4p9gw.js} +6 -6
  9. package/dist/index.js +1 -1
  10. package/dist/lib/manifest.d.ts +14 -0
  11. package/package.json +1 -1
package/README.md CHANGED
@@ -146,7 +146,7 @@ Agents are specialized subagents with pre-configured prompts and expertise. They
146
146
 
147
147
  ### Using Agents
148
148
 
149
- Agents are invoked via OpenCode's `@mention` syntax or `delegate_task`:
149
+ Agents are invoked via OpenCode's `@mention` syntax or `task`:
150
150
 
151
151
  ```
152
152
  @architecture-strategist Review the authentication refactoring in this PR
@@ -155,7 +155,7 @@ Agents are invoked via OpenCode's `@mention` syntax or `delegate_task`:
155
155
  Or programmatically in skills/commands:
156
156
 
157
157
  ```
158
- delegate_task(subagent_type="architecture-strategist", prompt="Review...")
158
+ task(subagent_type="architecture-strategist", prompt="Review...")
159
159
  ```
160
160
 
161
161
  ## Commands
@@ -387,7 +387,7 @@ See [`AGENTS.md`](./AGENTS.md) for detailed development guidelines, code style c
387
387
 
388
388
  ## Converting from Claude Code
389
389
 
390
- Migrating skills, agents, or commands from Claude Code (CEP) to Systematic? See the [Conversion Guide](https://fro.bot/systematic/guides/conversion-guide/) for field mappings and examples. Also available as [local Markdown](./docs/CONVERSION-GUIDE.md).
390
+ Migrating skills, agents, or commands from CEP or other Claude Code-format sources to Systematic? See the [Conversion Guide](https://fro.bot/systematic/guides/conversion-guide/) for field mappings and examples. Also available as [local Markdown](./docs/CONVERSION-GUIDE.md).
391
391
 
392
392
  ## References
393
393
 
package/commands/workflows/brainstorm.md CHANGED
@@ -33,7 +33,7 @@ Evaluate whether brainstorming is needed based on the feature description.
33
33
  - Constrained, well-defined scope
34
34
 
35
35
  **If requirements are already clear:**
36
- Use **AskUserQuestion tool** to suggest: "Your requirements seem detailed enough to proceed directly to planning. Should I run `/workflows:plan` instead, or would you like to explore the idea further?"
36
+ Use the **question tool** to suggest: "Your requirements seem detailed enough to proceed directly to planning. Should I run `/workflows:plan` instead, or would you like to explore the idea further?"
37
37
 
38
38
  ### Phase 1: Understand the Idea
39
39
 
@@ -41,13 +41,13 @@ Use **AskUserQuestion tool** to suggest: "Your requirements seem detailed enough
41
41
 
42
42
  Run a quick repo scan to understand existing patterns:
43
43
 
44
- - Task repo-research-analyst("Understand existing patterns related to: <feature_description>")
44
+ - task repo-research-analyst("Understand existing patterns related to: <feature_description>")
45
45
 
46
- Focus on: similar features, established patterns, CLAUDE.md guidance.
46
+ Focus on: similar features, established patterns, AGENTS.md guidance.
47
47
 
48
48
  #### 1.2 Collaborative Dialogue
49
49
 
50
- Use the **AskUserQuestion tool** to ask questions **one at a time**.
50
+ Use the **question tool** to ask questions **one at a time**.
51
51
 
52
52
  **Guidelines (see `brainstorming` skill for detailed techniques):**
53
53
  - Prefer multiple choice when natural options exist
@@ -68,7 +68,7 @@ For each approach, provide:
68
68
 
69
69
  Lead with your recommendation and explain why. Apply YAGNI—prefer simpler solutions.
70
70
 
71
- Use **AskUserQuestion tool** to ask which approach the user prefers.
71
+ Use the **question tool** to ask which approach the user prefers.
72
72
 
73
73
  ### Phase 3: Capture the Design
74
74
 
@@ -80,15 +80,24 @@ Ensure `docs/brainstorms/` directory exists before writing.
80
80
 
81
81
  ### Phase 4: Handoff
82
82
 
83
- Use **AskUserQuestion tool** to present next steps:
83
+ Use the **question tool** to present next steps:
84
84
 
85
85
  **Question:** "Brainstorm captured. What would you like to do next?"
86
86
 
87
87
  **Options:**
88
- 1. **Proceed to planning** - Run `/workflows:plan` (will auto-detect this brainstorm)
89
- 2. **Refine design further** - Continue exploring
88
+ 1. **Review and refine** - Improve the document through structured self-review
89
+ 2. **Proceed to planning** - Run `/workflows:plan` (will auto-detect this brainstorm)
90
90
  3. **Done for now** - Return later
91
91
 
92
+ **If user selects "Review and refine":**
93
+
94
+ Load the `document-review` skill and apply it to the brainstorm document.
95
+
96
+ When document-review returns "Review complete", present next steps:
97
+
98
+ 1. **Move to planning** - Continue to `/workflows:plan` with this document
99
+ 2. **Done for now** - Brainstorming complete. To start planning later: `/workflows:plan [document-path]`
100
+
92
101
  ## Output Summary
93
102
 
94
103
  When complete, display:
package/commands/workflows/compound.md CHANGED
@@ -21,53 +21,83 @@ Captures problem solutions while context is fresh, creating structured documenta
21
21
  /workflows:compound [brief context] # Provide additional context hint
22
22
  ```
23
23
 
24
- ## Execution Strategy: Parallel Subagents
25
-
26
- This command launches multiple specialized subagents IN PARALLEL to maximize efficiency:
27
-
28
- ### 1. **Context Analyzer** (Parallel)
29
- - Extracts conversation history
30
- - Identifies problem type, component, symptoms
31
- - Validates against CORA schema
32
- - Returns: YAML frontmatter skeleton
33
-
34
- ### 2. **Solution Extractor** (Parallel)
35
- - Analyzes all investigation steps
36
- - Identifies root cause
37
- - Extracts working solution with code examples
38
- - Returns: Solution content block
39
-
40
- ### 3. **Related Docs Finder** (Parallel)
41
- - Searches `docs/solutions/` for related documentation
42
- - Identifies cross-references and links
43
- - Finds related GitHub issues
44
- - Returns: Links and relationships
45
-
46
- ### 4. **Prevention Strategist** (Parallel)
47
- - Develops prevention strategies
48
- - Creates best practices guidance
49
- - Generates test cases if applicable
50
- - Returns: Prevention/testing content
51
-
52
- ### 5. **Category Classifier** (Parallel)
53
- - Determines optimal `docs/solutions/` category
54
- - Validates category against schema
55
- - Suggests filename based on slug
56
- - Returns: Final path and filename
57
-
58
- ### 6. **Documentation Writer** (Parallel)
59
- - Assembles complete markdown file
60
- - Validates YAML frontmatter
61
- - Formats content for readability
62
- - Creates the file in correct location
63
-
64
- ### 7. **Optional: Specialized Agent Invocation** (Post-Documentation)
65
- Based on problem type detected, automatically invoke applicable agents:
66
- - **performance_issue** → `performance-oracle`
67
- - **security_issue** → `security-sentinel`
68
- - **database_issue** → `data-integrity-guardian`
69
- - **test_failure** → `cora-test-reviewer`
70
- - Any code-heavy issue → `kieran-rails-reviewer` + `code-simplicity-reviewer`
24
+ ## Execution Strategy: Two-Phase Orchestration
25
+
26
+ <critical_requirement>
27
+ **Only ONE file gets written - the final documentation.**
28
+
29
+ Phase 1 subagents return TEXT DATA to the orchestrator. They must NOT use Write, Edit, or create any files. Only the orchestrator (Phase 2) writes the final documentation file.
30
+ </critical_requirement>
31
+
32
+ ### Phase 1: Parallel Research
33
+
34
+ <parallel_tasks>
35
+
36
+ Launch these subagents IN PARALLEL. Each returns text data to the orchestrator.
37
+
38
+ #### 1. **Context Analyzer**
39
+ - Extracts conversation history
40
+ - Identifies problem type, component, symptoms
41
+ - Validates against schema
42
+ - Returns: YAML frontmatter skeleton
43
+
44
+ #### 2. **Solution Extractor**
45
+ - Analyzes all investigation steps
46
+ - Identifies root cause
47
+ - Extracts working solution with code examples
48
+ - Returns: Solution content block
49
+
50
+ #### 3. **Related Docs Finder**
51
+ - Searches `docs/solutions/` for related documentation
52
+ - Identifies cross-references and links
53
+ - Finds related GitHub issues
54
+ - Returns: Links and relationships
55
+
56
+ #### 4. **Prevention Strategist**
57
+ - Develops prevention strategies
58
+ - Creates best practices guidance
59
+ - Generates test cases if applicable
60
+ - Returns: Prevention/testing content
61
+
62
+ #### 5. **Category Classifier**
63
+ - Determines optimal `docs/solutions/` category
64
+ - Validates category against schema
65
+ - Suggests filename based on slug
66
+ - Returns: Final path and filename
67
+
68
+ </parallel_tasks>
69
+
70
+ ### Phase 2: Assembly & Write
71
+
72
+ <sequential_tasks>
73
+
74
+ **WAIT for all Phase 1 subagents to complete before proceeding.**
75
+
76
+ The orchestrating agent (main conversation) performs these steps:
77
+
78
+ 1. Collect all text results from Phase 1 subagents
79
+ 2. Assemble complete markdown file from the collected pieces
80
+ 3. Validate YAML frontmatter against schema
81
+ 4. Create directory if needed: `mkdir -p docs/solutions/[category]/`
82
+ 5. Write the SINGLE final file: `docs/solutions/[category]/[filename].md`
83
+
84
+ </sequential_tasks>
85
+
86
+ ### Phase 3: Optional Enhancement
87
+
88
+ **WAIT for Phase 2 to complete before proceeding.**
89
+
90
+ <parallel_tasks>
91
+
92
+ Based on problem type, optionally invoke specialized agents to review the documentation:
93
+
94
+ - **performance_issue** → `performance-oracle`
95
+ - **security_issue** → `security-sentinel`
96
+ - **database_issue** → `data-integrity-guardian`
97
+ - **test_failure** → `cora-test-reviewer`
98
+ - Any code-heavy issue → `kieran-rails-reviewer` + `code-simplicity-reviewer`
99
+
100
+ </parallel_tasks>
71
101
 
72
102
  ## What It Captures
73
103
 
@@ -110,24 +140,29 @@ This command launches multiple specialized subagents IN PARALLEL to maximize eff
110
140
  - integration-issues/
111
141
  - logic-errors/
112
142
 
143
+ ## Common Mistakes to Avoid
144
+
145
+ | Wrong | Correct |
146
+ |----------|-----------:|
147
+ | Subagents write files like `context-analysis.md`, `solution-draft.md` | Subagents return text data; orchestrator writes one final file |
148
+ | Research and assembly run in parallel | Research completes, then assembly runs |
149
+ | Multiple files created during workflow | Single file: `docs/solutions/[category]/[filename].md` |
150
+
113
151
  ## Success Output
114
152
 
115
153
  ```
116
- ✓ Parallel documentation generation complete
117
-
118
- Primary Subagent Results:
119
- ✓ Context Analyzer: Identified performance_issue in brief_system
120
- ✓ Solution Extractor: Extracted 3 code fixes
121
- ✓ Related Docs Finder: Found 2 related issues
122
- ✓ Prevention Strategist: Generated test cases
123
- ✓ Category Classifier: docs/solutions/performance-issues/
124
- ✓ Documentation Writer: Created complete markdown
154
+ Subagent Results:
155
+ Context Analyzer: Identified performance_issue in brief_system
156
+ Solution Extractor: 3 code fixes
157
+ Related Docs Finder: 2 related issues
158
+ Prevention Strategist: Prevention strategies, test suggestions
159
+ Category Classifier: `performance-issues`
125
160
 
126
161
  Specialized Agent Reviews (Auto-Triggered):
127
- ✓ performance-oracle: Validated query optimization approach
128
- ✓ kieran-rails-reviewer: Code examples meet Rails standards
129
- ✓ code-simplicity-reviewer: Solution is appropriately minimal
130
- ✓ every-style-editor: Documentation style verified
162
+ performance-oracle: Validated query optimization approach
163
+ kieran-rails-reviewer: Code examples meet Rails standards
164
+ code-simplicity-reviewer: Solution is appropriately minimal
165
+ every-style-editor: Documentation style verified
131
166
 
132
167
  File created:
133
168
  - docs/solutions/performance-issues/n-plus-one-brief-generation.md
package/commands/workflows/plan.md CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  name: workflows:plan
3
3
  description: Transform feature descriptions into well-structured project plans following conventions
4
- argument-hint: "[feature description, bug report, or improvement idea]"
4
+ argument-hint: '[feature description, bug report, or improvement idea]'
5
5
  ---
6
6
 
7
7
  # Create a plan for a new feature or bug fix
@@ -43,11 +43,11 @@ ls -la docs/brainstorms/*.md 2>/dev/null | head -10
43
43
  5. Use brainstorm decisions as input to the research phase
44
44
 
45
45
  **If multiple brainstorms could match:**
46
- Use **AskUserQuestion tool** to ask which brainstorm to use, or whether to proceed without one.
46
+ Use **question tool** to ask which brainstorm to use, or whether to proceed without one.
47
47
 
48
48
  **If no brainstorm found (or not relevant), run idea refinement:**
49
49
 
50
- Refine the idea through collaborative dialogue using the **AskUserQuestion tool**:
50
+ Refine the idea through collaborative dialogue using the **question tool**:
51
51
 
52
52
  - Ask questions one at a time to understand the idea fully
53
53
  - Prefer multiple choice questions when natural options exist
@@ -74,11 +74,11 @@ First, I need to understand the project's conventions, existing patterns, and an
74
74
 
75
75
  Run these agents **in parallel** to gather local context:
76
76
 
77
- - Task repo-research-analyst(feature_description)
78
- - Task learnings-researcher(feature_description)
77
+ - task repo-research-analyst(feature_description)
78
+ - task learnings-researcher(feature_description)
79
79
 
80
80
  **What to look for:**
81
- - **Repo research:** existing patterns, CLAUDE.md guidance, technology familiarity, pattern consistency
81
+ - **Repo research:** existing patterns, AGENTS.md guidance, technology familiarity, pattern consistency
82
82
  - **Learnings:** documented solutions in `docs/solutions/` that might apply (gotchas, patterns, lessons learned)
83
83
 
84
84
  These findings inform the next step.
@@ -89,7 +89,7 @@ Based on signals from Step 0 and findings from Step 1, decide on external resear
89
89
 
90
90
  **High-risk topics → always research.** Security, payments, external APIs, data privacy. The cost of missing something is too high. This takes precedence over speed signals.
91
91
 
92
- **Strong local context → skip external research.** Codebase has good patterns, CLAUDE.md has guidance, user knows what they want. External research adds little value.
92
+ **Strong local context → skip external research.** Codebase has good patterns, AGENTS.md has guidance, user knows what they want. External research adds little value.
93
93
 
94
94
  **Uncertainty or unfamiliar territory → research.** User is exploring, codebase has no examples, new technology. External perspective is valuable.
95
95
 
@@ -105,8 +105,8 @@ Examples:
105
105
 
106
106
  Run these agents in parallel:
107
107
 
108
- - Task best-practices-researcher(feature_description)
109
- - Task framework-docs-researcher(feature_description)
108
+ - task best-practices-researcher(feature_description)
109
+ - task framework-docs-researcher(feature_description)
110
110
 
111
111
  ### 1.6. Consolidate Research
112
112
 
@@ -116,7 +116,7 @@ After all research steps complete, consolidate findings:
116
116
  - **Include relevant institutional learnings** from `docs/solutions/` (key insights, gotchas to avoid)
117
117
  - Note external documentation URLs and best practices (if external research was done)
118
118
  - List related issues or PRs discovered
119
- - Capture CLAUDE.md conventions
119
+ - Capture AGENTS.md conventions
120
120
 
121
121
  **Optional validation:** Briefly summarize findings and ask if anything looks off or missing before proceeding to planning.
122
122
 
@@ -150,7 +150,7 @@ Think like a product manager - what would make this issue clear and actionable?
150
150
 
151
151
  After planning the issue structure, run SpecFlow Analyzer to validate and refine the feature specification:
152
152
 
153
- - Task spec-flow-analyzer(feature_description, research_findings)
153
+ - task spec-flow-analyzer(feature_description, research_findings)
154
154
 
155
155
  **SpecFlow Analyzer Output:**
156
156
 
@@ -491,38 +491,36 @@ Examples:
491
491
 
492
492
  ## Post-Generation Options
493
493
 
494
- After writing the plan file, use the **AskUserQuestion tool** to present these options:
494
+ After writing the plan file, use the **question tool** to present these options:
495
495
 
496
496
  **Question:** "Plan ready at `docs/plans/YYYY-MM-DD-<type>-<name>-plan.md`. What would you like to do next?"
497
497
 
498
498
  **Options:**
499
499
  1. **Open plan in editor** - Open the plan file for review
500
500
  2. **Run `/deepen-plan`** - Enhance each section with parallel research agents (best practices, performance, UI)
501
- 3. **Run `/plan_review`** - Get feedback from reviewers (DHH, Kieran, Simplicity)
502
- 4. **Start `/workflows:work`** - Begin implementing this plan locally
503
- 5. **Start `/workflows:work` on remote** - Begin implementing in Claude Code on the web (use `&` to run in background)
501
+ 3. **Run `/technical_review`** - Technical feedback from code-focused reviewers (DHH, Kieran, Simplicity)
502
+ 4. **Review and refine** - Improve the document through structured self-review
503
+ 5. **Start `/workflows:work`** - Begin implementing this plan locally
504
504
  6. **Create Issue** - Create issue in project tracker (GitHub/Linear)
505
- 7. **Simplify** - Reduce detail level
506
505
 
507
506
  Based on selection:
508
507
  - **Open plan in editor** → Run `open docs/plans/<plan_filename>.md` to open the file in the user's default editor
509
508
  - **`/deepen-plan`** → Call the /deepen-plan command with the plan file path to enhance with research
510
- - **`/plan_review`** → Call the /plan_review command with the plan file path
509
+ - **`/technical_review`** → Call the /technical_review command with the plan file path
510
+ - **Review and refine** → Load `document-review` skill.
511
511
  - **`/workflows:work`** → Call the /workflows:work command with the plan file path
512
- - **`/workflows:work` on remote** → Run `/workflows:work docs/plans/<plan_filename>.md &` to start work in background for Claude Code web
513
512
  - **Create Issue** → See "Issue Creation" section below
514
- - **Simplify** → Ask "What should I simplify?" then regenerate simpler version
515
513
  - **Other** (automatically provided) → Accept free text for rework or specific changes
516
514
 
517
515
  **Note:** If running `/workflows:plan` with ultrathink enabled, automatically run `/deepen-plan` after plan creation for maximum depth and grounding.
518
516
 
519
- Loop back to options after Simplify or Other changes until user selects `/workflows:work` or `/plan_review`.
517
+ Loop back to options after Simplify or Other changes until user selects `/workflows:work` or `/technical_review`.
520
518
 
521
519
  ## Issue Creation
522
520
 
523
- When user selects "Create Issue", detect their project tracker from CLAUDE.md:
521
+ When user selects "Create Issue", detect their project tracker from AGENTS.md:
524
522
 
525
- 1. **Check for tracker preference** in user's CLAUDE.md (global or project):
523
+ 1. **Check for tracker preference** in user's AGENTS.md (global or project):
526
524
  - Look for `project_tracker: github` or `project_tracker: linear`
527
525
  - Or look for mentions of "GitHub Issues" or "Linear" in their workflow section
528
526
 
@@ -542,10 +540,10 @@ When user selects "Create Issue", detect their project tracker from CLAUDE.md:
542
540
 
543
541
  4. **If no tracker configured:**
544
542
  Ask user: "Which project tracker do you use? (GitHub/Linear/Other)"
545
- - Suggest adding `project_tracker: github` or `project_tracker: linear` to their CLAUDE.md
543
+ - Suggest adding `project_tracker: github` or `project_tracker: linear` to their AGENTS.md
546
544
 
547
545
  5. **After creation:**
548
546
  - Display the issue URL
549
- - Ask if they want to proceed to `/workflows:work` or `/plan_review`
547
+ - Ask if they want to proceed to `/workflows:work` or `/technical_review`
550
548
 
551
549
  NEVER CODE! Just research and write the plan.
package/commands/workflows/review.md CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  name: workflows:review
3
3
  description: Perform exhaustive code reviews using multi-agent analysis, ultra-thinking, and worktrees
4
- argument-hint: "[PR number, GitHub URL, branch name, or latest]"
4
+ argument-hint: '[PR number, GitHub URL, branch name, or latest]'
5
5
  ---
6
6
 
7
7
  # Review Command
@@ -31,7 +31,7 @@ argument-hint: "[PR number, GitHub URL, branch name, or latest]"
31
31
  First, I need to determine the review target type and set up the code for analysis.
32
32
  </thinking>
33
33
 
34
- #### Immediate Actions:
34
+ #### Immediate Actions
35
35
 
36
36
  <task_list>
37
37
 
@@ -48,29 +48,40 @@ Ensure that the code is ready for analysis (either in worktree or on current bra
48
48
 
49
49
  </task_list>
50
50
 
51
- #### Parallel Agents to review the PR:
51
+ #### Protected Artifacts
52
+
53
+ <protected_artifacts>
54
+ The following paths are systematic pipeline artifacts and must never be flagged for deletion, removal, or gitignore by any review agent:
55
+
56
+ - `docs/plans/*.md` — Plan files created by `/workflows:plan`. These are living documents that track implementation progress (checkboxes are checked off by `/workflows:work`).
57
+ - `docs/solutions/*.md` — Solution documents created during the pipeline.
58
+
59
+ If a review agent flags any file in these directories for cleanup or removal, discard that finding during synthesis. Do not create a todo for it.
60
+ </protected_artifacts>
61
+
62
+ #### Parallel Agents to review the PR
52
63
 
53
64
  <parallel_tasks>
54
65
 
55
66
  Run ALL or most of these agents at the same time:
56
67
 
57
- 1. Task kieran-rails-reviewer(PR content)
58
- 2. Task dhh-rails-reviewer(PR title)
59
- 3. If turbo is used: Task rails-turbo-expert(PR content)
60
- 4. Task git-history-analyzer(PR content)
61
- 5. Task dependency-detective(PR content)
62
- 6. Task pattern-recognition-specialist(PR content)
63
- 7. Task architecture-strategist(PR content)
64
- 8. Task code-philosopher(PR content)
65
- 9. Task security-sentinel(PR content)
66
- 10. Task performance-oracle(PR content)
67
- 11. Task devops-harmony-analyst(PR content)
68
- 12. Task data-integrity-guardian(PR content)
69
- 13. Task agent-native-reviewer(PR content) - Verify new features are agent-accessible
68
+ 1. task kieran-rails-reviewer(PR content)
69
+ 2. task dhh-rails-reviewer(PR title)
70
+ 3. If turbo is used: task rails-turbo-expert(PR content)
71
+ 4. task git-history-analyzer(PR content)
72
+ 5. task dependency-detective(PR content)
73
+ 6. task pattern-recognition-specialist(PR content)
74
+ 7. task architecture-strategist(PR content)
75
+ 8. task code-philosopher(PR content)
76
+ 9. task security-sentinel(PR content)
77
+ 10. task performance-oracle(PR content)
78
+ 11. task devops-harmony-analyst(PR content)
79
+ 12. task data-integrity-guardian(PR content)
80
+ 13. task agent-native-reviewer(PR content) - Verify new features are agent-accessible
70
81
 
71
82
  </parallel_tasks>
72
83
 
73
- #### Conditional Agents (Run if applicable):
84
+ #### Conditional Agents (Run if applicable)
74
85
 
75
86
  <conditional_agents>
76
87
 
@@ -78,8 +89,8 @@ These agents are run ONLY when the PR matches specific criteria. Check the PR fi
78
89
 
79
90
  **If PR contains database migrations (db/migrate/*.rb files) or data backfills:**
80
91
 
81
- 14. Task data-migration-expert(PR content) - Validates ID mappings match production, checks for swapped values, verifies rollback safety
82
- 15. Task deployment-verification-agent(PR content) - Creates Go/No-Go deployment checklist with SQL verification queries
92
+ 14. task data-migration-expert(PR content) - Validates ID mappings match production, checks for swapped values, verifies rollback safety
93
+ 15. task deployment-verification-agent(PR content) - Creates Go/No-Go deployment checklist with SQL verification queries
83
94
 
84
95
  **When to run migration agents:**
85
96
  - PR includes files matching `db/migrate/*.rb`
@@ -191,7 +202,7 @@ Complete system context map with component interactions
191
202
 
192
203
  ### 4. Simplification and Minimalism Review
193
204
 
194
- Run the Task code-simplicity-reviewer() to see if we can simplify the code.
205
+ Run the task code-simplicity-reviewer() to see if we can simplify the code.
195
206
 
196
207
  ### 5. Findings Synthesis and Todo Creation Using file-todos Skill
197
208
 
@@ -207,6 +218,7 @@ Remove duplicates, prioritize by severity and impact.
207
218
  <synthesis_tasks>
208
219
 
209
220
  - [ ] Collect findings from all parallel agents
221
+ - [ ] Discard any findings that recommend deleting or gitignoring files in `docs/plans/` or `docs/solutions/` (see Protected Artifacts above)
210
222
  - [ ] Categorize by type: security, performance, architecture, quality, etc.
211
223
  - [ ] Assign severity levels: 🔴 CRITICAL (P1), 🟡 IMPORTANT (P2), 🔵 NICE-TO-HAVE (P3)
212
224
  - [ ] Remove duplicate or overlapping findings
@@ -222,18 +234,18 @@ Remove duplicates, prioritize by severity and impact.
222
234
 
223
235
  **Option A: Direct File Creation (Fast)**
224
236
 
225
- - Create todo files directly using Write tool
237
+ - Create todo files directly using write tool
226
238
  - All findings in parallel for speed
227
- - Use standard template from `.claude/skills/file-todos/assets/todo-template.md`
239
+ - Use standard template from `.opencode/skills/file-todos/assets/todo-template.md`
228
240
  - Follow naming convention: `{issue_id}-pending-{priority}-{description}.md`
229
241
 
230
242
  **Option B: Sub-Agents in Parallel (Recommended for Scale)** For large PRs with 15+ findings, use sub-agents to create finding files in parallel:
231
243
 
232
244
  ```bash
233
245
  # Launch multiple finding-creator agents in parallel
234
- Task() - Create todos for first finding
235
- Task() - Create todos for second finding
236
- Task() - Create todos for third finding
246
+ task() - Create todos for first finding
247
+ task() - Create todos for second finding
248
+ task() - Create todos for third finding
237
249
  etc. for each finding.
238
250
  ```
239
251
 
@@ -272,7 +284,7 @@ Sub-agents can:
272
284
 
273
285
  The skill provides:
274
286
 
275
- - Template location: `.claude/skills/file-todos/assets/todo-template.md`
287
+ - Template location: `.opencode/skills/file-todos/assets/todo-template.md`
276
288
  - Naming convention: `{issue_id}-{status}-{priority}-{description}.md`
277
289
  - YAML frontmatter structure: status, priority, issue_id, tags, dependencies
278
290
  - All required sections: Problem Statement, Findings, Solutions, etc.
@@ -292,7 +304,7 @@ Sub-agents can:
292
304
  004-pending-p3-unused-parameter.md
293
305
  ```
294
306
 
295
- 5. Follow template structure from file-todos skill: `.claude/skills/file-todos/assets/todo-template.md`
307
+ 5. Follow template structure from file-todos skill: `.opencode/skills/file-todos/assets/todo-template.md`
296
308
 
297
309
  **Todo File Structure (from template):**
298
310
 
@@ -400,7 +412,7 @@ After creating all todo files, present comprehensive summary:
400
412
  - Update Work Log as you work
401
413
  - Commit todos: `git add todos/ && git commit -m "refactor: add code review findings"`
402
414
 
403
- ### Severity Breakdown:
415
+ ### Severity Breakdown
404
416
 
405
417
  **🔴 P1 (Critical - Blocks Merge):**
406
418
 
@@ -468,12 +480,12 @@ After presenting the Summary Report, offer appropriate testing based on project
468
480
 
469
481
  </offer_testing>
470
482
 
471
- #### If User Accepts Web Testing:
483
+ #### If User Accepts Web Testing
472
484
 
473
485
  Spawn a subagent to run browser tests (preserves main context):
474
486
 
475
487
  ```
476
- Task general-purpose("Run /test-browser for PR #[number]. Test all affected pages, check for console errors, handle failures by creating todos and fixing.")
488
+ task general-purpose("Run /test-browser for PR #[number]. Test all affected pages, check for console errors, handle failures by creating todos and fixing.")
477
489
  ```
478
490
 
479
491
  The subagent will:
@@ -487,12 +499,12 @@ The subagent will:
487
499
 
488
500
  **Standalone:** `/test-browser [PR number]`
489
501
 
490
- #### If User Accepts iOS Testing:
502
+ #### If User Accepts iOS Testing
491
503
 
492
504
  Spawn a subagent to run Xcode tests (preserves main context):
493
505
 
494
506
  ```
495
- Task general-purpose("Run /xcode-test for scheme [name]. Build for simulator, install, launch, take screenshots, check for crashes.")
507
+ task general-purpose("Run /xcode-test for scheme [name]. Build for simulator, install, launch, take screenshots, check for crashes.")
496
508
  ```
497
509
 
498
510
  The subagent will:
@@ -512,3 +524,4 @@ The subagent will:
512
524
 
513
525
  Any **🔴 P1 (CRITICAL)** findings must be addressed before merging the PR. Present these prominently and ensure they're resolved before accepting the PR.
514
526
  ```
527
+
package/commands/workflows/work.md CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  name: workflows:work
3
3
  description: Execute work plans efficiently while maintaining quality and finishing features
4
- argument-hint: "[plan file, specification, or todo file path]"
4
+ argument-hint: '[plan file, specification, or todo file path]'
5
5
  ---
6
6
 
7
7
  # Work Plan Execution Command
@@ -73,7 +73,7 @@ This command takes a work document (plan, specification, or todo file) and execu
73
73
  - You plan to switch between branches frequently
74
74
 
75
75
  3. **Create Todo List**
76
- - Use TodoWrite to break plan into actionable tasks
76
+ - Use todowrite to break plan into actionable tasks
77
77
  - Include dependencies between tasks
78
78
  - Prioritize based on what needs to be done first
79
79
  - Include testing and quality check tasks
@@ -87,18 +87,18 @@ This command takes a work document (plan, specification, or todo file) and execu
87
87
 
88
88
  ```
89
89
  while (tasks remain):
90
- - Mark task as in_progress in TodoWrite
90
+ - Mark task as in_progress in todowrite
91
91
  - Read any referenced files from the plan
92
92
  - Look for similar patterns in codebase
93
93
  - Implement following existing conventions
94
94
  - Write tests for new functionality
95
95
  - Run tests after changes
96
- - Mark task as completed in TodoWrite
96
+ - Mark task as completed in todowrite
97
97
  - Mark off the corresponding checkbox in the plan file ([ ] → [x])
98
98
  - Evaluate for incremental commit (see below)
99
99
  ```
100
100
 
101
- **IMPORTANT**: Always update the original plan document by checking off completed items. Use the Edit tool to change `- [ ]` to `- [x]` for each task you finish. This keeps the plan as a living document showing progress and ensures no checkboxes are left unchecked.
101
+ **IMPORTANT**: Always update the original plan document by checking off completed items. Use the edit tool to change `- [ ]` to `- [x]` for each task you finish. This keeps the plan as a living document showing progress and ensures no checkboxes are left unchecked.
102
102
 
103
103
  2. **Incremental Commits**
104
104
 
@@ -134,7 +134,7 @@ This command takes a work document (plan, specification, or todo file) and execu
134
134
  - The plan should reference similar code - read those files first
135
135
  - Match naming conventions exactly
136
136
  - Reuse existing components where possible
137
- - Follow project coding standards (see CLAUDE.md)
137
+ - Follow project coding standards (see AGENTS.md)
138
138
  - When in doubt, grep for similar implementations
139
139
 
140
140
  4. **Test Continuously**
@@ -154,7 +154,7 @@ This command takes a work document (plan, specification, or todo file) and execu
154
154
  - Repeat until implementation matches design
155
155
 
156
156
  6. **Track Progress**
157
- - Keep TodoWrite updated as you complete tasks
157
+ - Keep todowrite updated as you complete tasks
158
158
  - Note any blockers or unexpected discoveries
159
159
  - Create new tasks if scope expands
160
160
  - Keep user informed of major milestones
@@ -169,7 +169,7 @@ This command takes a work document (plan, specification, or todo file) and execu
169
169
  # Run full test suite (use project's test command)
170
170
  # Examples: bin/rails test, npm test, pytest, go test, etc.
171
171
 
172
- # Run linting (per CLAUDE.md)
172
+ # Run linting (per AGENTS.md)
173
173
  # Use linting-agent before pushing to origin
174
174
  ```
175
175
 
@@ -181,19 +181,19 @@ This command takes a work document (plan, specification, or todo file) and execu
181
181
  - **kieran-rails-reviewer**: Verify Rails conventions (Rails projects)
182
182
  - **performance-oracle**: Check for performance issues
183
183
  - **security-sentinel**: Scan for security vulnerabilities
184
- - **cora-test-reviewer**: Review test quality (CORA projects)
184
+ - **cora-test-reviewer**: Review test quality (Rails projects with comprehensive test coverage)
185
185
 
186
- Run reviewers in parallel with Task tool:
186
+ Run reviewers in parallel with task tool:
187
187
 
188
188
  ```
189
- Task(code-simplicity-reviewer): "Review changes for simplicity"
190
- Task(kieran-rails-reviewer): "Check Rails conventions"
189
+ task(code-simplicity-reviewer): "Review changes for simplicity"
190
+ task(kieran-rails-reviewer): "Check Rails conventions"
191
191
  ```
192
192
 
193
193
  Present findings to user and address critical issues.
194
194
 
195
195
  3. **Final Validation**
196
- - All TodoWrite tasks marked completed
196
+ - All todowrite tasks marked completed
197
197
  - All tests pass
198
198
  - Linting passes
199
199
  - Code follows existing patterns
@@ -215,9 +215,9 @@ This command takes a work document (plan, specification, or todo file) and execu
215
215
 
216
216
  Brief explanation if needed.
217
217
 
218
- 🤖 Generated with [Claude Code](https://claude.com/claude-code)
218
+ 🤖 Generated with [OpenCode](https://opencode.ai)
219
219
 
220
- Co-Authored-By: Claude <noreply@anthropic.com>
220
+ Co-Authored-By: OpenCode <noreply@opencode.ai>
221
221
  EOF
222
222
  )"
223
223
  ```
@@ -279,7 +279,7 @@ This command takes a work document (plan, specification, or todo file) and execu
279
279
 
280
280
  ---
281
281
 
282
- [![Compound Engineered](https://img.shields.io/badge/Compound-Engineered-6366f1)](https://github.com/EveryInc/compound-engineering-plugin) 🤖 Generated with [Claude Code](https://claude.com/claude-code)
282
+ [![Systematic](https://img.shields.io/badge/Systematic-Engineered-6366f1)](https://github.com/marcusrbrown/systematic) 🤖 Generated with [OpenCode](https://opencode.ai)
283
283
  EOF
284
284
  )"
285
285
  ```
@@ -292,6 +292,78 @@ This command takes a work document (plan, specification, or todo file) and execu
292
292
 
293
293
  ---
294
294
 
295
+ ## Swarm Mode (Optional)
296
+
297
+ > **Note:** Swarm mode coordination primitives (`Teammate` API — team creation, shutdown signals, cleanup) are not yet available in OpenCode. The concepts below are aspirational. For current parallel execution, use the `task` tool with multiple background subagents, or see the `dispatching-parallel-agents` skill.
298
+
299
+ For complex plans with multiple independent workstreams, enable swarm mode for parallel execution with coordinated agents.
300
+
301
+ ### When to Use Swarm Mode
302
+
303
+ | Use Swarm Mode when... | Use Standard Mode when... |
304
+ |------------------------|---------------------------|
305
+ | Plan has 5+ independent tasks | Plan is linear/sequential |
306
+ | Multiple specialists needed (review + test + implement) | Single-focus work |
307
+ | Want maximum parallelism | Simpler mental model preferred |
308
+ | Large feature with clear phases | Small feature or bug fix |
309
+
310
+ ### Enabling Swarm Mode
311
+
312
+ To trigger swarm execution, say:
313
+
314
+ > "Make a Task list and launch an army of agent swarm subagents to build the plan"
315
+
316
+ Or explicitly request: "Use swarm mode for this work"
317
+
318
+ ### Swarm Workflow
319
+
320
+ When swarm mode is enabled, the workflow changes:
321
+
322
+ 1. **Create Team**
323
+ ```
324
+ Teammate({ operation: "spawnTeam", team_name: "work-{timestamp}" })
325
+ ```
326
+
327
+ 2. **Create Task List with Dependencies**
328
+ - Parse plan into TaskCreate items
329
+ - Set up blockedBy relationships for sequential dependencies
330
+ - Independent tasks have no blockers (can run in parallel)
331
+
332
+ 3. **Spawn Specialized Teammates**
333
+ ```
334
+ Task({
335
+ team_name: "work-{timestamp}",
336
+ name: "implementer",
337
+ subagent_type: "general-purpose",
338
+ prompt: "Claim implementation tasks, execute, mark complete",
339
+ run_in_background: true
340
+ })
341
+
342
+ Task({
343
+ team_name: "work-{timestamp}",
344
+ name: "tester",
345
+ subagent_type: "general-purpose",
346
+ prompt: "Claim testing tasks, run tests, mark complete",
347
+ run_in_background: true
348
+ })
349
+ ```
350
+
351
+ 4. **Coordinate and Monitor**
352
+ - Team lead monitors task completion
353
+ - Spawn additional workers as phases unblock
354
+ - Handle plan approval if required
355
+
356
+ 5. **Cleanup**
357
+ ```
358
+ Teammate({ operation: "requestShutdown", target_agent_id: "implementer" })
359
+ Teammate({ operation: "requestShutdown", target_agent_id: "tester" })
360
+ Teammate({ operation: "cleanup" })
361
+ ```
362
+
363
+ See the `orchestrating-swarms` skill for detailed swarm patterns and best practices.
364
+
365
+ ---
366
+
295
367
  ## Key Principles
296
368
 
297
369
  ### Start Fast, Execute Faster
@@ -330,7 +402,7 @@ This command takes a work document (plan, specification, or todo file) and execu
330
402
  Before creating PR, verify:
331
403
 
332
404
  - [ ] All clarifying questions asked and answered
333
- - [ ] All TodoWrite tasks marked completed
405
+ - [ ] All todowrite tasks marked completed
334
406
  - [ ] Tests pass (run project's test command)
335
407
  - [ ] Linting passes (use linting-agent)
336
408
  - [ ] Code follows existing patterns
@@ -338,7 +410,7 @@ Before creating PR, verify:
338
410
  - [ ] Before/after screenshots captured and uploaded (for UI changes)
339
411
  - [ ] Commit messages follow conventional format
340
412
  - [ ] PR description includes summary, testing notes, and screenshots
341
- - [ ] PR description includes Compound Engineered badge
413
+ - [ ] PR description includes Systematic badge
342
414
 
343
415
  ## When to Use Reviewer Agents
344
416
 
@@ -358,6 +430,6 @@ For most features: tests + linting + following patterns is sufficient.
358
430
  - **Skipping clarifying questions** - Ask now, not after building wrong thing
359
431
  - **Ignoring plan references** - The plan has links for a reason
360
432
  - **Testing at the end** - Test continuously or suffer later
361
- - **Forgetting TodoWrite** - Track progress or lose track of what's done
433
+ - **Forgetting todowrite** - Track progress or lose track of what's done
362
434
  - **80% done syndrome** - Finish the feature, don't move on early
363
435
  - **Over-reviewing simple changes** - Save reviewer agents for complex work
package/dist/cli.js CHANGED
@@ -6,7 +6,7 @@ import {
6
6
  findCommandsInDir,
7
7
  findSkillsInDir,
8
8
  getConfigPaths
9
- } from "./index-0ftaxvrt.js";
9
+ } from "./index-bky4p9gw.js";
10
10
 
11
11
  // src/cli.ts
12
12
  import fs from "fs";
package/dist/{index-0ftaxvrt.js → index-bky4p9gw.js} RENAMED
@@ -322,11 +322,11 @@ import fs3 from "fs";
322
322
  var CONVERTER_VERSION = 2;
323
323
  var cache = new Map;
324
324
  var TOOL_MAPPINGS = [
325
- [/\bTask\s+tool\b/gi, "delegate_task tool"],
326
- [/\bTask\s+([\w-]+)\s*:/g, "delegate_task $1:"],
327
- [/\bTask\s+([\w-]+)\s*\(/g, "delegate_task $1("],
328
- [/\bTask\s*\(/g, "delegate_task("],
329
- [/\bTask\b(?=\s+to\s+\w)/g, "delegate_task"],
325
+ [/\bTask\s+tool\b/gi, "task tool"],
326
+ [/\bTask\s+([\w-]+)\s*:/g, "task $1:"],
327
+ [/\bTask\s+([\w-]+)\s*\(/g, "task $1("],
328
+ [/\bTask\s*\(/g, "task("],
329
+ [/\bTask\b(?=\s+to\s+\w)/g, "task"],
330
330
  [/\bTodoWrite\b/g, "todowrite"],
331
331
  [/\bAskUserQuestion\b/g, "question"],
332
332
  [/\bWebSearch\b/g, "google_search"],
@@ -349,7 +349,7 @@ var PATH_REPLACEMENTS = [
349
349
  [/compound-engineering:/g, "systematic:"]
350
350
  ];
351
351
  var TOOL_NAME_MAP = {
352
- task: "delegate_task",
352
+ task: "task",
353
353
  todowrite: "todowrite",
354
354
  askuserquestion: "question",
355
355
  websearch: "google_search",
package/dist/index.js CHANGED
@@ -8,7 +8,7 @@ import {
8
8
  findSkillsInDir,
9
9
  loadConfig,
10
10
  parseFrontmatter
11
- } from "./index-0ftaxvrt.js";
11
+ } from "./index-bky4p9gw.js";
12
12
 
13
13
  // src/index.ts
14
14
  import fs3 from "fs";
package/dist/lib/manifest.d.ts CHANGED
@@ -3,12 +3,26 @@ export interface ManifestSource {
3
3
  branch: string;
4
4
  url: string;
5
5
  }
6
+ export interface ManifestRewrite {
7
+ field: string;
8
+ reason: string;
9
+ original?: string;
10
+ }
11
+ export interface ManualOverride {
12
+ field: string;
13
+ reason: string;
14
+ original?: string;
15
+ overridden_at: string;
16
+ }
6
17
  export interface ManifestDefinition {
7
18
  source: string;
8
19
  upstream_path: string;
9
20
  upstream_commit: string;
10
21
  synced_at: string;
11
22
  notes: string;
23
+ upstream_content_hash?: string;
24
+ rewrites?: ManifestRewrite[];
25
+ manual_overrides?: ManualOverride[];
12
26
  }
13
27
  export interface SyncManifest {
14
28
  $schema?: string;
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@fro.bot/systematic",
3
- "version": "1.12.0",
3
+ "version": "1.13.0",
4
4
  "description": "Structured engineering workflows for OpenCode",
5
5
  "type": "module",
6
6
  "homepage": "https://fro.bot/systematic",