claude-context-plugin 1.0.0 → 1.2.0
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.
- package/.claude-plugin/plugin.json +23 -23
- package/README.md +104 -104
- package/agents/context-engineer.md +67 -67
- package/package.json +32 -32
- package/skills/implement/SKILL.md +71 -71
- package/skills/plan/SKILL.md +66 -66
- package/skills/research/SKILL.md +43 -43
- package/skills/validate/SKILL.md +62 -62
- package/skills/verify-docs/SKILL.md +77 -77
|
@@ -1,23 +1,23 @@
|
|
|
1
|
-
{
|
|
2
|
-
"name": "claude-context-engineering",
|
|
3
|
-
"version": "1.0.0",
|
|
4
|
-
"displayName": "Context Engineering",
|
|
5
|
-
"description": "RPI workflow and context optimization for Claude Code - Research, Plan, Implement with file:line precision",
|
|
6
|
-
"author": {
|
|
7
|
-
"name": "SireJeff"
|
|
8
|
-
},
|
|
9
|
-
"license": "MIT",
|
|
10
|
-
"repository": {
|
|
11
|
-
"type": "git",
|
|
12
|
-
"url": "https://github.com/SireJeff/claude-context-engineering-template"
|
|
13
|
-
},
|
|
14
|
-
"keywords": [
|
|
15
|
-
"context-engineering",
|
|
16
|
-
"rpi-workflow",
|
|
17
|
-
"documentation",
|
|
18
|
-
"productivity"
|
|
19
|
-
],
|
|
20
|
-
"claudeCode": {
|
|
21
|
-
"minVersion": "1.0.33"
|
|
22
|
-
}
|
|
23
|
-
}
|
|
1
|
+
{
|
|
2
|
+
"name": "claude-context-engineering",
|
|
3
|
+
"version": "1.0.0",
|
|
4
|
+
"displayName": "Context Engineering",
|
|
5
|
+
"description": "RPI workflow and context optimization for Claude Code - Research, Plan, Implement with file:line precision",
|
|
6
|
+
"author": {
|
|
7
|
+
"name": "SireJeff"
|
|
8
|
+
},
|
|
9
|
+
"license": "MIT",
|
|
10
|
+
"repository": {
|
|
11
|
+
"type": "git",
|
|
12
|
+
"url": "https://github.com/SireJeff/claude-context-engineering-template"
|
|
13
|
+
},
|
|
14
|
+
"keywords": [
|
|
15
|
+
"context-engineering",
|
|
16
|
+
"rpi-workflow",
|
|
17
|
+
"documentation",
|
|
18
|
+
"productivity"
|
|
19
|
+
],
|
|
20
|
+
"claudeCode": {
|
|
21
|
+
"minVersion": "1.0.33"
|
|
22
|
+
}
|
|
23
|
+
}
|
package/README.md
CHANGED
|
@@ -1,104 +1,104 @@
|
|
|
1
|
-
# claude-context-plugin
|
|
2
|
-
|
|
3
|
-
Claude Code plugin for context engineering - RPI workflow and documentation optimization.
|
|
4
|
-
|
|
5
|
-
## Installation
|
|
6
|
-
|
|
7
|
-
### Via Plugin Manager
|
|
8
|
-
|
|
9
|
-
```bash
|
|
10
|
-
/plugin install claude-context-engineering
|
|
11
|
-
```
|
|
12
|
-
|
|
13
|
-
### Via CLI
|
|
14
|
-
|
|
15
|
-
```bash
|
|
16
|
-
npx claude-plugins install claude-context-engineering
|
|
17
|
-
```
|
|
18
|
-
|
|
19
|
-
## Skills
|
|
20
|
-
|
|
21
|
-
### /context-eng:research
|
|
22
|
-
|
|
23
|
-
Systematic codebase exploration for feature/fix analysis.
|
|
24
|
-
|
|
25
|
-
```
|
|
26
|
-
/context-eng:research user-authentication
|
|
27
|
-
```
|
|
28
|
-
|
|
29
|
-
Creates a research document with:
|
|
30
|
-
- Relevant files and line numbers
|
|
31
|
-
- Call chain analysis
|
|
32
|
-
- Dependency mapping
|
|
33
|
-
- Test coverage gaps
|
|
34
|
-
|
|
35
|
-
### /context-eng:plan
|
|
36
|
-
|
|
37
|
-
Create implementation blueprint with file:line precision.
|
|
38
|
-
|
|
39
|
-
```
|
|
40
|
-
/context-eng:plan user-authentication
|
|
41
|
-
```
|
|
42
|
-
|
|
43
|
-
Creates a plan document with:
|
|
44
|
-
- Exact file:line modifications
|
|
45
|
-
- Risk assessment
|
|
46
|
-
- Test strategy
|
|
47
|
-
- Rollback plan
|
|
48
|
-
|
|
49
|
-
### /context-eng:implement
|
|
50
|
-
|
|
51
|
-
Execute plan with atomic changes and continuous testing.
|
|
52
|
-
|
|
53
|
-
```
|
|
54
|
-
/context-eng:implement user-authentication
|
|
55
|
-
```
|
|
56
|
-
|
|
57
|
-
Follows the golden rule: ONE CHANGE → ONE TEST → ONE COMMIT
|
|
58
|
-
|
|
59
|
-
### /context-eng:validate
|
|
60
|
-
|
|
61
|
-
Run complete validation suite.
|
|
62
|
-
|
|
63
|
-
```
|
|
64
|
-
/context-eng:validate
|
|
65
|
-
```
|
|
66
|
-
|
|
67
|
-
Checks:
|
|
68
|
-
- Directory structure
|
|
69
|
-
- JSON schemas
|
|
70
|
-
- Markdown links
|
|
71
|
-
- Placeholder completion
|
|
72
|
-
- Line number accuracy
|
|
73
|
-
|
|
74
|
-
### /context-eng:verify-docs
|
|
75
|
-
|
|
76
|
-
Verify documentation line numbers match code.
|
|
77
|
-
|
|
78
|
-
```
|
|
79
|
-
/context-eng:verify-docs path/to/file.py
|
|
80
|
-
```
|
|
81
|
-
|
|
82
|
-
## Agents
|
|
83
|
-
|
|
84
|
-
### @context-eng:context-engineer
|
|
85
|
-
|
|
86
|
-
Initialize and maintain context engineering documentation.
|
|
87
|
-
|
|
88
|
-
```
|
|
89
|
-
@context-eng:context-engineer "Discover workflows for this codebase"
|
|
90
|
-
```
|
|
91
|
-
|
|
92
|
-
## Requirements
|
|
93
|
-
|
|
94
|
-
- Claude Code v1.0.33+
|
|
95
|
-
- Context engineering setup (use `npx create-claude-context` first)
|
|
96
|
-
|
|
97
|
-
## License
|
|
98
|
-
|
|
99
|
-
MIT
|
|
100
|
-
|
|
101
|
-
## Links
|
|
102
|
-
|
|
103
|
-
- [GitHub](https://github.com/SireJeff/claude-context-engineering-template)
|
|
104
|
-
- [create-claude-context](https://www.npmjs.com/package/create-claude-context)
|
|
1
|
+
# claude-context-plugin
|
|
2
|
+
|
|
3
|
+
Claude Code plugin for context engineering - RPI workflow and documentation optimization.
|
|
4
|
+
|
|
5
|
+
## Installation
|
|
6
|
+
|
|
7
|
+
### Via Plugin Manager
|
|
8
|
+
|
|
9
|
+
```bash
|
|
10
|
+
/plugin install claude-context-engineering
|
|
11
|
+
```
|
|
12
|
+
|
|
13
|
+
### Via CLI
|
|
14
|
+
|
|
15
|
+
```bash
|
|
16
|
+
npx claude-plugins install claude-context-engineering
|
|
17
|
+
```
|
|
18
|
+
|
|
19
|
+
## Skills
|
|
20
|
+
|
|
21
|
+
### /context-eng:research
|
|
22
|
+
|
|
23
|
+
Systematic codebase exploration for feature/fix analysis.
|
|
24
|
+
|
|
25
|
+
```
|
|
26
|
+
/context-eng:research user-authentication
|
|
27
|
+
```
|
|
28
|
+
|
|
29
|
+
Creates a research document with:
|
|
30
|
+
- Relevant files and line numbers
|
|
31
|
+
- Call chain analysis
|
|
32
|
+
- Dependency mapping
|
|
33
|
+
- Test coverage gaps
|
|
34
|
+
|
|
35
|
+
### /context-eng:plan
|
|
36
|
+
|
|
37
|
+
Create implementation blueprint with file:line precision.
|
|
38
|
+
|
|
39
|
+
```
|
|
40
|
+
/context-eng:plan user-authentication
|
|
41
|
+
```
|
|
42
|
+
|
|
43
|
+
Creates a plan document with:
|
|
44
|
+
- Exact file:line modifications
|
|
45
|
+
- Risk assessment
|
|
46
|
+
- Test strategy
|
|
47
|
+
- Rollback plan
|
|
48
|
+
|
|
49
|
+
### /context-eng:implement
|
|
50
|
+
|
|
51
|
+
Execute plan with atomic changes and continuous testing.
|
|
52
|
+
|
|
53
|
+
```
|
|
54
|
+
/context-eng:implement user-authentication
|
|
55
|
+
```
|
|
56
|
+
|
|
57
|
+
Follows the golden rule: ONE CHANGE → ONE TEST → ONE COMMIT
|
|
58
|
+
|
|
59
|
+
### /context-eng:validate
|
|
60
|
+
|
|
61
|
+
Run complete validation suite.
|
|
62
|
+
|
|
63
|
+
```
|
|
64
|
+
/context-eng:validate
|
|
65
|
+
```
|
|
66
|
+
|
|
67
|
+
Checks:
|
|
68
|
+
- Directory structure
|
|
69
|
+
- JSON schemas
|
|
70
|
+
- Markdown links
|
|
71
|
+
- Placeholder completion
|
|
72
|
+
- Line number accuracy
|
|
73
|
+
|
|
74
|
+
### /context-eng:verify-docs
|
|
75
|
+
|
|
76
|
+
Verify documentation line numbers match code.
|
|
77
|
+
|
|
78
|
+
```
|
|
79
|
+
/context-eng:verify-docs path/to/file.py
|
|
80
|
+
```
|
|
81
|
+
|
|
82
|
+
## Agents
|
|
83
|
+
|
|
84
|
+
### @context-eng:context-engineer
|
|
85
|
+
|
|
86
|
+
Initialize and maintain context engineering documentation.
|
|
87
|
+
|
|
88
|
+
```
|
|
89
|
+
@context-eng:context-engineer "Discover workflows for this codebase"
|
|
90
|
+
```
|
|
91
|
+
|
|
92
|
+
## Requirements
|
|
93
|
+
|
|
94
|
+
- Claude Code v1.0.33+
|
|
95
|
+
- Context engineering setup (use `npx create-claude-context` first)
|
|
96
|
+
|
|
97
|
+
## License
|
|
98
|
+
|
|
99
|
+
MIT
|
|
100
|
+
|
|
101
|
+
## Links
|
|
102
|
+
|
|
103
|
+
- [GitHub](https://github.com/SireJeff/claude-context-engineering-template)
|
|
104
|
+
- [create-claude-context](https://www.npmjs.com/package/create-claude-context)
|
|
@@ -1,67 +1,67 @@
|
|
|
1
|
-
---
|
|
2
|
-
name: context-engineer
|
|
3
|
-
description: Initialize and maintain context engineering documentation for any codebase
|
|
4
|
-
capabilities:
|
|
5
|
-
- Codebase analysis and workflow discovery
|
|
6
|
-
- Documentation generation with file:line references
|
|
7
|
-
- Index creation and maintenance
|
|
8
|
-
- Tech stack detection
|
|
9
|
-
---
|
|
10
|
-
|
|
11
|
-
# Context Engineer Agent
|
|
12
|
-
|
|
13
|
-
I specialize in setting up and maintaining context engineering systems for codebases.
|
|
14
|
-
|
|
15
|
-
## Primary Functions
|
|
16
|
-
|
|
17
|
-
### 1. Initial Setup
|
|
18
|
-
When asked to initialize or set up context engineering:
|
|
19
|
-
- Analyze codebase structure
|
|
20
|
-
- Detect technology stack
|
|
21
|
-
- Discover major workflows (8-15)
|
|
22
|
-
- Create workflow documentation with file:line references
|
|
23
|
-
- Build index hierarchy
|
|
24
|
-
|
|
25
|
-
### 2. Workflow Discovery
|
|
26
|
-
When asked to discover or find workflows:
|
|
27
|
-
- Scan for entry points (API routes, CLI commands, event handlers)
|
|
28
|
-
- Trace call chains 3 levels deep
|
|
29
|
-
- Document in `.claude/context/workflows/`
|
|
30
|
-
- Update `WORKFLOW_INDEX.md`
|
|
31
|
-
|
|
32
|
-
### 3. Re-indexing
|
|
33
|
-
When asked to re-scan or update indexes:
|
|
34
|
-
- Find new or changed files
|
|
35
|
-
- Update line numbers in documentation
|
|
36
|
-
- Add new workflows if discovered
|
|
37
|
-
- Validate existing references
|
|
38
|
-
|
|
39
|
-
### 4. Documentation Maintenance
|
|
40
|
-
When documentation drift is detected:
|
|
41
|
-
- Update line numbers
|
|
42
|
-
- Fix broken references
|
|
43
|
-
- Add missing workflows
|
|
44
|
-
- Archive outdated docs
|
|
45
|
-
|
|
46
|
-
## Output Format
|
|
47
|
-
|
|
48
|
-
All documentation follows these standards:
|
|
49
|
-
- File references: `path/to/file.py:123`
|
|
50
|
-
- Line ranges: `file.py:100-150`
|
|
51
|
-
- Markdown tables for structured data
|
|
52
|
-
- Mermaid diagrams for flows
|
|
53
|
-
|
|
54
|
-
## Context Budget
|
|
55
|
-
|
|
56
|
-
- Initial setup: 40-50% utilization
|
|
57
|
-
- Re-indexing: 20-30% utilization
|
|
58
|
-
- Maintenance: 15-20% utilization
|
|
59
|
-
|
|
60
|
-
## Example Invocations
|
|
61
|
-
|
|
62
|
-
```
|
|
63
|
-
@context-engineer "Initialize context engineering for this repository"
|
|
64
|
-
@context-engineer "Discover workflows for the authentication module"
|
|
65
|
-
@context-engineer "Re-scan and update indexes for recent changes"
|
|
66
|
-
@context-engineer "Update line numbers in workflow documentation"
|
|
67
|
-
```
|
|
1
|
+
---
|
|
2
|
+
name: context-engineer
|
|
3
|
+
description: Initialize and maintain context engineering documentation for any codebase
|
|
4
|
+
capabilities:
|
|
5
|
+
- Codebase analysis and workflow discovery
|
|
6
|
+
- Documentation generation with file:line references
|
|
7
|
+
- Index creation and maintenance
|
|
8
|
+
- Tech stack detection
|
|
9
|
+
---
|
|
10
|
+
|
|
11
|
+
# Context Engineer Agent
|
|
12
|
+
|
|
13
|
+
I specialize in setting up and maintaining context engineering systems for codebases.
|
|
14
|
+
|
|
15
|
+
## Primary Functions
|
|
16
|
+
|
|
17
|
+
### 1. Initial Setup
|
|
18
|
+
When asked to initialize or set up context engineering:
|
|
19
|
+
- Analyze codebase structure
|
|
20
|
+
- Detect technology stack
|
|
21
|
+
- Discover major workflows (8-15)
|
|
22
|
+
- Create workflow documentation with file:line references
|
|
23
|
+
- Build index hierarchy
|
|
24
|
+
|
|
25
|
+
### 2. Workflow Discovery
|
|
26
|
+
When asked to discover or find workflows:
|
|
27
|
+
- Scan for entry points (API routes, CLI commands, event handlers)
|
|
28
|
+
- Trace call chains 3 levels deep
|
|
29
|
+
- Document in `.claude/context/workflows/`
|
|
30
|
+
- Update `WORKFLOW_INDEX.md`
|
|
31
|
+
|
|
32
|
+
### 3. Re-indexing
|
|
33
|
+
When asked to re-scan or update indexes:
|
|
34
|
+
- Find new or changed files
|
|
35
|
+
- Update line numbers in documentation
|
|
36
|
+
- Add new workflows if discovered
|
|
37
|
+
- Validate existing references
|
|
38
|
+
|
|
39
|
+
### 4. Documentation Maintenance
|
|
40
|
+
When documentation drift is detected:
|
|
41
|
+
- Update line numbers
|
|
42
|
+
- Fix broken references
|
|
43
|
+
- Add missing workflows
|
|
44
|
+
- Archive outdated docs
|
|
45
|
+
|
|
46
|
+
## Output Format
|
|
47
|
+
|
|
48
|
+
All documentation follows these standards:
|
|
49
|
+
- File references: `path/to/file.py:123`
|
|
50
|
+
- Line ranges: `file.py:100-150`
|
|
51
|
+
- Markdown tables for structured data
|
|
52
|
+
- Mermaid diagrams for flows
|
|
53
|
+
|
|
54
|
+
## Context Budget
|
|
55
|
+
|
|
56
|
+
- Initial setup: 40-50% utilization
|
|
57
|
+
- Re-indexing: 20-30% utilization
|
|
58
|
+
- Maintenance: 15-20% utilization
|
|
59
|
+
|
|
60
|
+
## Example Invocations
|
|
61
|
+
|
|
62
|
+
```
|
|
63
|
+
@context-engineer "Initialize context engineering for this repository"
|
|
64
|
+
@context-engineer "Discover workflows for the authentication module"
|
|
65
|
+
@context-engineer "Re-scan and update indexes for recent changes"
|
|
66
|
+
@context-engineer "Update line numbers in workflow documentation"
|
|
67
|
+
```
|
package/package.json
CHANGED
|
@@ -1,32 +1,32 @@
|
|
|
1
|
-
{
|
|
2
|
-
"name": "claude-context-plugin",
|
|
3
|
-
"version": "1.
|
|
4
|
-
"description": "Claude Code plugin for context engineering - RPI workflow and documentation optimization",
|
|
5
|
-
"keywords": [
|
|
6
|
-
"claude-code",
|
|
7
|
-
"plugin",
|
|
8
|
-
"context-engineering",
|
|
9
|
-
"rpi-workflow",
|
|
10
|
-
"documentation"
|
|
11
|
-
],
|
|
12
|
-
"author": "SireJeff",
|
|
13
|
-
"license": "MIT",
|
|
14
|
-
"repository": {
|
|
15
|
-
"type": "git",
|
|
16
|
-
"url": "git+https://github.com/SireJeff/claude-context-engineering-template.git",
|
|
17
|
-
"directory": "packages/claude-context-plugin"
|
|
18
|
-
},
|
|
19
|
-
"bugs": {
|
|
20
|
-
"url": "https://github.com/SireJeff/claude-context-engineering-template/issues"
|
|
21
|
-
},
|
|
22
|
-
"homepage": "https://github.com/SireJeff/claude-context-engineering-template#readme",
|
|
23
|
-
"files": [
|
|
24
|
-
".claude-plugin/",
|
|
25
|
-
"skills/",
|
|
26
|
-
"agents/",
|
|
27
|
-
"README.md"
|
|
28
|
-
],
|
|
29
|
-
"peerDependencies": {
|
|
30
|
-
"claude-code": ">=1.0.33"
|
|
31
|
-
}
|
|
32
|
-
}
|
|
1
|
+
{
|
|
2
|
+
"name": "claude-context-plugin",
|
|
3
|
+
"version": "1.2.0",
|
|
4
|
+
"description": "Claude Code plugin for context engineering - RPI workflow and documentation optimization",
|
|
5
|
+
"keywords": [
|
|
6
|
+
"claude-code",
|
|
7
|
+
"plugin",
|
|
8
|
+
"context-engineering",
|
|
9
|
+
"rpi-workflow",
|
|
10
|
+
"documentation"
|
|
11
|
+
],
|
|
12
|
+
"author": "SireJeff",
|
|
13
|
+
"license": "MIT",
|
|
14
|
+
"repository": {
|
|
15
|
+
"type": "git",
|
|
16
|
+
"url": "git+https://github.com/SireJeff/claude-context-engineering-template.git",
|
|
17
|
+
"directory": "packages/claude-context-plugin"
|
|
18
|
+
},
|
|
19
|
+
"bugs": {
|
|
20
|
+
"url": "https://github.com/SireJeff/claude-context-engineering-template/issues"
|
|
21
|
+
},
|
|
22
|
+
"homepage": "https://github.com/SireJeff/claude-context-engineering-template#readme",
|
|
23
|
+
"files": [
|
|
24
|
+
".claude-plugin/",
|
|
25
|
+
"skills/",
|
|
26
|
+
"agents/",
|
|
27
|
+
"README.md"
|
|
28
|
+
],
|
|
29
|
+
"peerDependencies": {
|
|
30
|
+
"claude-code": ">=1.0.33"
|
|
31
|
+
}
|
|
32
|
+
}
|
|
@@ -1,71 +1,71 @@
|
|
|
1
|
-
---
|
|
2
|
-
description: RPI Implement Phase - Execute plan with atomic changes and continuous testing
|
|
3
|
-
---
|
|
4
|
-
|
|
5
|
-
# Context Engineering: Implement Phase
|
|
6
|
-
|
|
7
|
-
When invoked, execute the approved implementation plan:
|
|
8
|
-
|
|
9
|
-
## Prerequisites
|
|
10
|
-
- Approved plan at `.claude/plans/active/[feature]_plan.md`
|
|
11
|
-
- If not found, run `/context-eng:plan $ARGUMENTS` first
|
|
12
|
-
|
|
13
|
-
## Golden Rule
|
|
14
|
-
|
|
15
|
-
```
|
|
16
|
-
ONE CHANGE → ONE TEST → ONE COMMIT
|
|
17
|
-
```
|
|
18
|
-
|
|
19
|
-
## Process
|
|
20
|
-
|
|
21
|
-
1. **Load Plan Document**
|
|
22
|
-
- Read `.claude/plans/active/[feature]_plan.md`
|
|
23
|
-
- Verify plan status is APPROVED
|
|
24
|
-
|
|
25
|
-
2. **For Each Step in Plan:**
|
|
26
|
-
a. Make the single, atomic change
|
|
27
|
-
b. Run the specified test immediately
|
|
28
|
-
c. If test passes: commit with descriptive message
|
|
29
|
-
d. If test fails: STOP, investigate, fix before proceeding
|
|
30
|
-
e. Update documentation if needed
|
|
31
|
-
|
|
32
|
-
3. **Documentation Updates (MANDATORY)**
|
|
33
|
-
- Check `CODE_TO_WORKFLOW_MAP.md` for affected workflows
|
|
34
|
-
- Update workflow files with new line numbers
|
|
35
|
-
- Update function signatures if changed
|
|
36
|
-
|
|
37
|
-
4. **Context Reset (Every 3 Steps)**
|
|
38
|
-
- Update progress checklist in plan
|
|
39
|
-
- Re-read plan document
|
|
40
|
-
- Verify scope alignment
|
|
41
|
-
- Compact if >35% utilization
|
|
42
|
-
|
|
43
|
-
5. **Finalize**
|
|
44
|
-
- Move plan to `.claude/plans/completed/`
|
|
45
|
-
- Move research to `.claude/research/completed/`
|
|
46
|
-
- Run `/context-eng:validate` to verify
|
|
47
|
-
|
|
48
|
-
## Error Handling
|
|
49
|
-
|
|
50
|
-
| Error Type | Response |
|
|
51
|
-
|------------|----------|
|
|
52
|
-
| Syntax Error | STOP. Fix immediately. |
|
|
53
|
-
| Import Error | Check file paths, verify imports. |
|
|
54
|
-
| Test Failure | Do NOT add more code. Investigate first. |
|
|
55
|
-
| 3+ Failures | STOP. Start new session. |
|
|
56
|
-
|
|
57
|
-
## Context Budget
|
|
58
|
-
- Plan: 15k tokens
|
|
59
|
-
- Active code: 30k tokens
|
|
60
|
-
- Test results: 15k tokens
|
|
61
|
-
- Total: 60k tokens (30%)
|
|
62
|
-
|
|
63
|
-
## Commit Format
|
|
64
|
-
```
|
|
65
|
-
feat/fix/refactor(scope): description
|
|
66
|
-
|
|
67
|
-
- Detail 1
|
|
68
|
-
- Detail 2
|
|
69
|
-
|
|
70
|
-
Implements: [feature]
|
|
71
|
-
```
|
|
1
|
+
---
|
|
2
|
+
description: RPI Implement Phase - Execute plan with atomic changes and continuous testing
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
# Context Engineering: Implement Phase
|
|
6
|
+
|
|
7
|
+
When invoked, execute the approved implementation plan:
|
|
8
|
+
|
|
9
|
+
## Prerequisites
|
|
10
|
+
- Approved plan at `.claude/plans/active/[feature]_plan.md`
|
|
11
|
+
- If not found, run `/context-eng:plan $ARGUMENTS` first
|
|
12
|
+
|
|
13
|
+
## Golden Rule
|
|
14
|
+
|
|
15
|
+
```
|
|
16
|
+
ONE CHANGE → ONE TEST → ONE COMMIT
|
|
17
|
+
```
|
|
18
|
+
|
|
19
|
+
## Process
|
|
20
|
+
|
|
21
|
+
1. **Load Plan Document**
|
|
22
|
+
- Read `.claude/plans/active/[feature]_plan.md`
|
|
23
|
+
- Verify plan status is APPROVED
|
|
24
|
+
|
|
25
|
+
2. **For Each Step in Plan:**
|
|
26
|
+
a. Make the single, atomic change
|
|
27
|
+
b. Run the specified test immediately
|
|
28
|
+
c. If test passes: commit with descriptive message
|
|
29
|
+
d. If test fails: STOP, investigate, fix before proceeding
|
|
30
|
+
e. Update documentation if needed
|
|
31
|
+
|
|
32
|
+
3. **Documentation Updates (MANDATORY)**
|
|
33
|
+
- Check `CODE_TO_WORKFLOW_MAP.md` for affected workflows
|
|
34
|
+
- Update workflow files with new line numbers
|
|
35
|
+
- Update function signatures if changed
|
|
36
|
+
|
|
37
|
+
4. **Context Reset (Every 3 Steps)**
|
|
38
|
+
- Update progress checklist in plan
|
|
39
|
+
- Re-read plan document
|
|
40
|
+
- Verify scope alignment
|
|
41
|
+
- Compact if >35% utilization
|
|
42
|
+
|
|
43
|
+
5. **Finalize**
|
|
44
|
+
- Move plan to `.claude/plans/completed/`
|
|
45
|
+
- Move research to `.claude/research/completed/`
|
|
46
|
+
- Run `/context-eng:validate` to verify
|
|
47
|
+
|
|
48
|
+
## Error Handling
|
|
49
|
+
|
|
50
|
+
| Error Type | Response |
|
|
51
|
+
|------------|----------|
|
|
52
|
+
| Syntax Error | STOP. Fix immediately. |
|
|
53
|
+
| Import Error | Check file paths, verify imports. |
|
|
54
|
+
| Test Failure | Do NOT add more code. Investigate first. |
|
|
55
|
+
| 3+ Failures | STOP. Start new session. |
|
|
56
|
+
|
|
57
|
+
## Context Budget
|
|
58
|
+
- Plan: 15k tokens
|
|
59
|
+
- Active code: 30k tokens
|
|
60
|
+
- Test results: 15k tokens
|
|
61
|
+
- Total: 60k tokens (30%)
|
|
62
|
+
|
|
63
|
+
## Commit Format
|
|
64
|
+
```
|
|
65
|
+
feat/fix/refactor(scope): description
|
|
66
|
+
|
|
67
|
+
- Detail 1
|
|
68
|
+
- Detail 2
|
|
69
|
+
|
|
70
|
+
Implements: [feature]
|
|
71
|
+
```
|
package/skills/plan/SKILL.md
CHANGED
|
@@ -1,66 +1,66 @@
|
|
|
1
|
-
---
|
|
2
|
-
description: RPI Plan Phase - Create implementation blueprint with file:line precision
|
|
3
|
-
---
|
|
4
|
-
|
|
5
|
-
# Context Engineering: Plan Phase
|
|
6
|
-
|
|
7
|
-
When invoked, create a detailed implementation plan:
|
|
8
|
-
|
|
9
|
-
## Prerequisites
|
|
10
|
-
- Research document exists at `.claude/research/active/[feature]_research.md`
|
|
11
|
-
- If not found, run `/context-eng:research $ARGUMENTS` first
|
|
12
|
-
|
|
13
|
-
## Process
|
|
14
|
-
|
|
15
|
-
1. **Load Research Document**
|
|
16
|
-
- Read the research document for $ARGUMENTS
|
|
17
|
-
- Extract relevant files and line numbers
|
|
18
|
-
|
|
19
|
-
2. **Design Implementation**
|
|
20
|
-
- For each change required:
|
|
21
|
-
- Specify exact file and line numbers
|
|
22
|
-
- Show current code snippet
|
|
23
|
-
- Show proposed change
|
|
24
|
-
- Assign risk level (LOW/MEDIUM/HIGH)
|
|
25
|
-
|
|
26
|
-
3. **Plan Test Strategy**
|
|
27
|
-
- Identify which tests to run after each change
|
|
28
|
-
- Note any new tests needed
|
|
29
|
-
|
|
30
|
-
4. **Document Rollback Plan**
|
|
31
|
-
- Safe commit to revert to
|
|
32
|
-
- Commands to undo changes
|
|
33
|
-
|
|
34
|
-
5. **Create Plan Document**
|
|
35
|
-
- Save to `.claude/plans/active/[feature]_plan.md`
|
|
36
|
-
- Include verification checklist
|
|
37
|
-
|
|
38
|
-
## Plan Format
|
|
39
|
-
|
|
40
|
-
```markdown
|
|
41
|
-
# Implementation Plan: [Feature]
|
|
42
|
-
|
|
43
|
-
## Files to Modify
|
|
44
|
-
| File | Lines | Change | Risk |
|
|
45
|
-
|------|-------|--------|------|
|
|
46
|
-
|
|
47
|
-
## Step-by-Step
|
|
48
|
-
### Step 1: [Action]
|
|
49
|
-
**File:** path/to/file.ext
|
|
50
|
-
**Lines:** X-Y
|
|
51
|
-
**Current:** [code block]
|
|
52
|
-
**Proposed:** [code block]
|
|
53
|
-
**Test:** [command]
|
|
54
|
-
|
|
55
|
-
## Rollback
|
|
56
|
-
- Safe commit: [hash]
|
|
57
|
-
- Command: git revert HEAD
|
|
58
|
-
```
|
|
59
|
-
|
|
60
|
-
## Context Budget
|
|
61
|
-
- Research doc: 20k tokens
|
|
62
|
-
- Plan creation: 15k tokens
|
|
63
|
-
- Total: 35k tokens (17.5%)
|
|
64
|
-
|
|
65
|
-
## Next Step
|
|
66
|
-
After approval, run `/context-eng:implement $ARGUMENTS`
|
|
1
|
+
---
|
|
2
|
+
description: RPI Plan Phase - Create implementation blueprint with file:line precision
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
# Context Engineering: Plan Phase
|
|
6
|
+
|
|
7
|
+
When invoked, create a detailed implementation plan:
|
|
8
|
+
|
|
9
|
+
## Prerequisites
|
|
10
|
+
- Research document exists at `.claude/research/active/[feature]_research.md`
|
|
11
|
+
- If not found, run `/context-eng:research $ARGUMENTS` first
|
|
12
|
+
|
|
13
|
+
## Process
|
|
14
|
+
|
|
15
|
+
1. **Load Research Document**
|
|
16
|
+
- Read the research document for $ARGUMENTS
|
|
17
|
+
- Extract relevant files and line numbers
|
|
18
|
+
|
|
19
|
+
2. **Design Implementation**
|
|
20
|
+
- For each change required:
|
|
21
|
+
- Specify exact file and line numbers
|
|
22
|
+
- Show current code snippet
|
|
23
|
+
- Show proposed change
|
|
24
|
+
- Assign risk level (LOW/MEDIUM/HIGH)
|
|
25
|
+
|
|
26
|
+
3. **Plan Test Strategy**
|
|
27
|
+
- Identify which tests to run after each change
|
|
28
|
+
- Note any new tests needed
|
|
29
|
+
|
|
30
|
+
4. **Document Rollback Plan**
|
|
31
|
+
- Safe commit to revert to
|
|
32
|
+
- Commands to undo changes
|
|
33
|
+
|
|
34
|
+
5. **Create Plan Document**
|
|
35
|
+
- Save to `.claude/plans/active/[feature]_plan.md`
|
|
36
|
+
- Include verification checklist
|
|
37
|
+
|
|
38
|
+
## Plan Format
|
|
39
|
+
|
|
40
|
+
```markdown
|
|
41
|
+
# Implementation Plan: [Feature]
|
|
42
|
+
|
|
43
|
+
## Files to Modify
|
|
44
|
+
| File | Lines | Change | Risk |
|
|
45
|
+
|------|-------|--------|------|
|
|
46
|
+
|
|
47
|
+
## Step-by-Step
|
|
48
|
+
### Step 1: [Action]
|
|
49
|
+
**File:** path/to/file.ext
|
|
50
|
+
**Lines:** X-Y
|
|
51
|
+
**Current:** [code block]
|
|
52
|
+
**Proposed:** [code block]
|
|
53
|
+
**Test:** [command]
|
|
54
|
+
|
|
55
|
+
## Rollback
|
|
56
|
+
- Safe commit: [hash]
|
|
57
|
+
- Command: git revert HEAD
|
|
58
|
+
```
|
|
59
|
+
|
|
60
|
+
## Context Budget
|
|
61
|
+
- Research doc: 20k tokens
|
|
62
|
+
- Plan creation: 15k tokens
|
|
63
|
+
- Total: 35k tokens (17.5%)
|
|
64
|
+
|
|
65
|
+
## Next Step
|
|
66
|
+
After approval, run `/context-eng:implement $ARGUMENTS`
|
package/skills/research/SKILL.md
CHANGED
|
@@ -1,43 +1,43 @@
|
|
|
1
|
-
---
|
|
2
|
-
description: RPI Research Phase - Systematic codebase exploration for feature/fix analysis
|
|
3
|
-
---
|
|
4
|
-
|
|
5
|
-
# Context Engineering: Research Phase
|
|
6
|
-
|
|
7
|
-
When invoked, perform systematic codebase exploration:
|
|
8
|
-
|
|
9
|
-
## Process
|
|
10
|
-
|
|
11
|
-
1. **Initialize Research Document**
|
|
12
|
-
- Create `.claude/research/active/[feature]_research.md`
|
|
13
|
-
- Use template from `.claude/research/RESEARCH_TEMPLATE.md`
|
|
14
|
-
|
|
15
|
-
2. **Entry Point Discovery**
|
|
16
|
-
- Search for API routes, CLI commands, event handlers related to $ARGUMENTS
|
|
17
|
-
- Use parallel exploration for efficiency (up to 3 agents)
|
|
18
|
-
|
|
19
|
-
3. **Call Chain Tracing**
|
|
20
|
-
- Trace 3 levels deep from entry points
|
|
21
|
-
- Record exact file:line references for each function
|
|
22
|
-
|
|
23
|
-
4. **Dependency Mapping**
|
|
24
|
-
- Internal: services, models, utilities used
|
|
25
|
-
- External: APIs, databases, third-party services
|
|
26
|
-
|
|
27
|
-
5. **Test Coverage Analysis**
|
|
28
|
-
- Find existing tests for affected code
|
|
29
|
-
- Identify coverage gaps
|
|
30
|
-
|
|
31
|
-
6. **Generate Summary**
|
|
32
|
-
- 150-word summary for Plan phase
|
|
33
|
-
- List 3-20 relevant files with line numbers
|
|
34
|
-
|
|
35
|
-
## Context Budget
|
|
36
|
-
- Target: 25% of 200k tokens (50k)
|
|
37
|
-
- Final output: ~20k tokens (research doc only)
|
|
38
|
-
|
|
39
|
-
## Output
|
|
40
|
-
Research document saved to `.claude/research/active/`
|
|
41
|
-
|
|
42
|
-
## Next Step
|
|
43
|
-
After completion, run `/context-eng:plan $ARGUMENTS`
|
|
1
|
+
---
|
|
2
|
+
description: RPI Research Phase - Systematic codebase exploration for feature/fix analysis
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
# Context Engineering: Research Phase
|
|
6
|
+
|
|
7
|
+
When invoked, perform systematic codebase exploration:
|
|
8
|
+
|
|
9
|
+
## Process
|
|
10
|
+
|
|
11
|
+
1. **Initialize Research Document**
|
|
12
|
+
- Create `.claude/research/active/[feature]_research.md`
|
|
13
|
+
- Use template from `.claude/research/RESEARCH_TEMPLATE.md`
|
|
14
|
+
|
|
15
|
+
2. **Entry Point Discovery**
|
|
16
|
+
- Search for API routes, CLI commands, event handlers related to $ARGUMENTS
|
|
17
|
+
- Use parallel exploration for efficiency (up to 3 agents)
|
|
18
|
+
|
|
19
|
+
3. **Call Chain Tracing**
|
|
20
|
+
- Trace 3 levels deep from entry points
|
|
21
|
+
- Record exact file:line references for each function
|
|
22
|
+
|
|
23
|
+
4. **Dependency Mapping**
|
|
24
|
+
- Internal: services, models, utilities used
|
|
25
|
+
- External: APIs, databases, third-party services
|
|
26
|
+
|
|
27
|
+
5. **Test Coverage Analysis**
|
|
28
|
+
- Find existing tests for affected code
|
|
29
|
+
- Identify coverage gaps
|
|
30
|
+
|
|
31
|
+
6. **Generate Summary**
|
|
32
|
+
- 150-word summary for Plan phase
|
|
33
|
+
- List 3-20 relevant files with line numbers
|
|
34
|
+
|
|
35
|
+
## Context Budget
|
|
36
|
+
- Target: 25% of 200k tokens (50k)
|
|
37
|
+
- Final output: ~20k tokens (research doc only)
|
|
38
|
+
|
|
39
|
+
## Output
|
|
40
|
+
Research document saved to `.claude/research/active/`
|
|
41
|
+
|
|
42
|
+
## Next Step
|
|
43
|
+
After completion, run `/context-eng:plan $ARGUMENTS`
|
package/skills/validate/SKILL.md
CHANGED
|
@@ -1,62 +1,62 @@
|
|
|
1
|
-
---
|
|
2
|
-
description: Run complete validation suite on context engineering documentation
|
|
3
|
-
---
|
|
4
|
-
|
|
5
|
-
# Context Engineering: Validation
|
|
6
|
-
|
|
7
|
-
Run comprehensive validation on the context engineering setup.
|
|
8
|
-
|
|
9
|
-
## Validation Checks
|
|
10
|
-
|
|
11
|
-
### 1. Structure Validation
|
|
12
|
-
- All required directories exist in `.claude/`
|
|
13
|
-
- Required files are present
|
|
14
|
-
- Settings file is valid JSON
|
|
15
|
-
|
|
16
|
-
### 2. Schema Validation
|
|
17
|
-
- JSON files validate against their schemas
|
|
18
|
-
- Frontmatter in markdown files is valid
|
|
19
|
-
|
|
20
|
-
### 3. Link Validation
|
|
21
|
-
- All internal markdown links resolve
|
|
22
|
-
- No broken references to files or anchors
|
|
23
|
-
|
|
24
|
-
### 4. Placeholder Detection
|
|
25
|
-
- Find remaining `{{PLACEHOLDER}}` values
|
|
26
|
-
- Report files that need manual completion
|
|
27
|
-
|
|
28
|
-
### 5. Line Number Accuracy
|
|
29
|
-
- Sample referenced line numbers in documentation
|
|
30
|
-
- Verify they match actual code
|
|
31
|
-
- Target: ≥60% accuracy
|
|
32
|
-
|
|
33
|
-
## Output
|
|
34
|
-
|
|
35
|
-
```
|
|
36
|
-
Validation Results
|
|
37
|
-
==================
|
|
38
|
-
✓ Structure: PASS (12/12 directories)
|
|
39
|
-
✓ Schemas: PASS (5/5 files valid)
|
|
40
|
-
✓ Links: PASS (45/45 links resolve)
|
|
41
|
-
⚠ Placeholders: WARN (3 remaining in 2 files)
|
|
42
|
-
✓ Line Accuracy: PASS (78% accurate)
|
|
43
|
-
|
|
44
|
-
Overall: PASS with warnings
|
|
45
|
-
```
|
|
46
|
-
|
|
47
|
-
## Usage
|
|
48
|
-
|
|
49
|
-
Run validation:
|
|
50
|
-
- Full suite: `/context-eng:validate`
|
|
51
|
-
- Specific check: `/context-eng:validate --links`
|
|
52
|
-
- With fix: `/context-eng:validate --fix`
|
|
53
|
-
|
|
54
|
-
## Thresholds
|
|
55
|
-
|
|
56
|
-
| Check | Pass | Warn | Fail |
|
|
57
|
-
|-------|------|------|------|
|
|
58
|
-
| Structure | 100% | - | <100% |
|
|
59
|
-
| Schemas | 100% | - | <100% |
|
|
60
|
-
| Links | 100% | <100% | <90% |
|
|
61
|
-
| Placeholders | 0 | 1-5 | >5 |
|
|
62
|
-
| Line Accuracy | ≥60% | 50-60% | <50% |
|
|
1
|
+
---
|
|
2
|
+
description: Run complete validation suite on context engineering documentation
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
# Context Engineering: Validation
|
|
6
|
+
|
|
7
|
+
Run comprehensive validation on the context engineering setup.
|
|
8
|
+
|
|
9
|
+
## Validation Checks
|
|
10
|
+
|
|
11
|
+
### 1. Structure Validation
|
|
12
|
+
- All required directories exist in `.claude/`
|
|
13
|
+
- Required files are present
|
|
14
|
+
- Settings file is valid JSON
|
|
15
|
+
|
|
16
|
+
### 2. Schema Validation
|
|
17
|
+
- JSON files validate against their schemas
|
|
18
|
+
- Frontmatter in markdown files is valid
|
|
19
|
+
|
|
20
|
+
### 3. Link Validation
|
|
21
|
+
- All internal markdown links resolve
|
|
22
|
+
- No broken references to files or anchors
|
|
23
|
+
|
|
24
|
+
### 4. Placeholder Detection
|
|
25
|
+
- Find remaining `{{PLACEHOLDER}}` values
|
|
26
|
+
- Report files that need manual completion
|
|
27
|
+
|
|
28
|
+
### 5. Line Number Accuracy
|
|
29
|
+
- Sample referenced line numbers in documentation
|
|
30
|
+
- Verify they match actual code
|
|
31
|
+
- Target: ≥60% accuracy
|
|
32
|
+
|
|
33
|
+
## Output
|
|
34
|
+
|
|
35
|
+
```
|
|
36
|
+
Validation Results
|
|
37
|
+
==================
|
|
38
|
+
✓ Structure: PASS (12/12 directories)
|
|
39
|
+
✓ Schemas: PASS (5/5 files valid)
|
|
40
|
+
✓ Links: PASS (45/45 links resolve)
|
|
41
|
+
⚠ Placeholders: WARN (3 remaining in 2 files)
|
|
42
|
+
✓ Line Accuracy: PASS (78% accurate)
|
|
43
|
+
|
|
44
|
+
Overall: PASS with warnings
|
|
45
|
+
```
|
|
46
|
+
|
|
47
|
+
## Usage
|
|
48
|
+
|
|
49
|
+
Run validation:
|
|
50
|
+
- Full suite: `/context-eng:validate`
|
|
51
|
+
- Specific check: `/context-eng:validate --links`
|
|
52
|
+
- With fix: `/context-eng:validate --fix`
|
|
53
|
+
|
|
54
|
+
## Thresholds
|
|
55
|
+
|
|
56
|
+
| Check | Pass | Warn | Fail |
|
|
57
|
+
|-------|------|------|------|
|
|
58
|
+
| Structure | 100% | - | <100% |
|
|
59
|
+
| Schemas | 100% | - | <100% |
|
|
60
|
+
| Links | 100% | <100% | <90% |
|
|
61
|
+
| Placeholders | 0 | 1-5 | >5 |
|
|
62
|
+
| Line Accuracy | ≥60% | 50-60% | <50% |
|
|
@@ -1,77 +1,77 @@
|
|
|
1
|
-
---
|
|
2
|
-
description: Verify documentation line numbers match current code state
|
|
3
|
-
---
|
|
4
|
-
|
|
5
|
-
# Context Engineering: Verify Documentation
|
|
6
|
-
|
|
7
|
-
Check if documentation line numbers still match the actual code.
|
|
8
|
-
|
|
9
|
-
## Purpose
|
|
10
|
-
|
|
11
|
-
After code changes, line numbers in documentation may drift. This skill:
|
|
12
|
-
1. Finds all `file:line` references in `.claude/` documentation
|
|
13
|
-
2. Verifies the referenced lines exist and match described content
|
|
14
|
-
3. Reports accuracy and suggests fixes
|
|
15
|
-
|
|
16
|
-
## Process
|
|
17
|
-
|
|
18
|
-
1. **Scan Documentation**
|
|
19
|
-
- Find all patterns like `file.py:123` or `[file.py:123]`
|
|
20
|
-
- Extract file path and line number
|
|
21
|
-
|
|
22
|
-
2. **Verify References**
|
|
23
|
-
- Check if file exists
|
|
24
|
-
- Check if line number is within file bounds
|
|
25
|
-
- Sample content to verify it matches description
|
|
26
|
-
|
|
27
|
-
3. **Calculate Accuracy**
|
|
28
|
-
- Total references found
|
|
29
|
-
- Valid references
|
|
30
|
-
- Accuracy percentage
|
|
31
|
-
|
|
32
|
-
4. **Generate Report**
|
|
33
|
-
```
|
|
34
|
-
Documentation Accuracy Report
|
|
35
|
-
=============================
|
|
36
|
-
Total References: 156
|
|
37
|
-
Valid: 142
|
|
38
|
-
Invalid: 14
|
|
39
|
-
Accuracy: 91%
|
|
40
|
-
|
|
41
|
-
Invalid References:
|
|
42
|
-
- context/workflows/auth.md:45 → src/auth.py:230 (line moved to 245)
|
|
43
|
-
- context/workflows/api.md:12 → src/api.py:100 (file renamed)
|
|
44
|
-
```
|
|
45
|
-
|
|
46
|
-
## Usage
|
|
47
|
-
|
|
48
|
-
Check specific file:
|
|
49
|
-
```
|
|
50
|
-
/context-eng:verify-docs path/to/changed/file.py
|
|
51
|
-
```
|
|
52
|
-
|
|
53
|
-
Check all documentation:
|
|
54
|
-
```
|
|
55
|
-
/context-eng:verify-docs
|
|
56
|
-
```
|
|
57
|
-
|
|
58
|
-
Auto-fix (update line numbers):
|
|
59
|
-
```
|
|
60
|
-
/context-eng:verify-docs --fix
|
|
61
|
-
```
|
|
62
|
-
|
|
63
|
-
## Target Accuracy
|
|
64
|
-
|
|
65
|
-
| Level | Accuracy | Action |
|
|
66
|
-
|-------|----------|--------|
|
|
67
|
-
| Excellent | ≥80% | No action needed |
|
|
68
|
-
| Good | 60-80% | Update when convenient |
|
|
69
|
-
| Warning | 40-60% | Update soon |
|
|
70
|
-
| Critical | <40% | Update immediately |
|
|
71
|
-
|
|
72
|
-
## When to Run
|
|
73
|
-
|
|
74
|
-
- After refactoring
|
|
75
|
-
- After adding/removing significant code
|
|
76
|
-
- Before creating a PR
|
|
77
|
-
- Weekly maintenance
|
|
1
|
+
---
|
|
2
|
+
description: Verify documentation line numbers match current code state
|
|
3
|
+
---
|
|
4
|
+
|
|
5
|
+
# Context Engineering: Verify Documentation
|
|
6
|
+
|
|
7
|
+
Check if documentation line numbers still match the actual code.
|
|
8
|
+
|
|
9
|
+
## Purpose
|
|
10
|
+
|
|
11
|
+
After code changes, line numbers in documentation may drift. This skill:
|
|
12
|
+
1. Finds all `file:line` references in `.claude/` documentation
|
|
13
|
+
2. Verifies the referenced lines exist and match described content
|
|
14
|
+
3. Reports accuracy and suggests fixes
|
|
15
|
+
|
|
16
|
+
## Process
|
|
17
|
+
|
|
18
|
+
1. **Scan Documentation**
|
|
19
|
+
- Find all patterns like `file.py:123` or `[file.py:123]`
|
|
20
|
+
- Extract file path and line number
|
|
21
|
+
|
|
22
|
+
2. **Verify References**
|
|
23
|
+
- Check if file exists
|
|
24
|
+
- Check if line number is within file bounds
|
|
25
|
+
- Sample content to verify it matches description
|
|
26
|
+
|
|
27
|
+
3. **Calculate Accuracy**
|
|
28
|
+
- Total references found
|
|
29
|
+
- Valid references
|
|
30
|
+
- Accuracy percentage
|
|
31
|
+
|
|
32
|
+
4. **Generate Report**
|
|
33
|
+
```
|
|
34
|
+
Documentation Accuracy Report
|
|
35
|
+
=============================
|
|
36
|
+
Total References: 156
|
|
37
|
+
Valid: 142
|
|
38
|
+
Invalid: 14
|
|
39
|
+
Accuracy: 91%
|
|
40
|
+
|
|
41
|
+
Invalid References:
|
|
42
|
+
- context/workflows/auth.md:45 → src/auth.py:230 (line moved to 245)
|
|
43
|
+
- context/workflows/api.md:12 → src/api.py:100 (file renamed)
|
|
44
|
+
```
|
|
45
|
+
|
|
46
|
+
## Usage
|
|
47
|
+
|
|
48
|
+
Check specific file:
|
|
49
|
+
```
|
|
50
|
+
/context-eng:verify-docs path/to/changed/file.py
|
|
51
|
+
```
|
|
52
|
+
|
|
53
|
+
Check all documentation:
|
|
54
|
+
```
|
|
55
|
+
/context-eng:verify-docs
|
|
56
|
+
```
|
|
57
|
+
|
|
58
|
+
Auto-fix (update line numbers):
|
|
59
|
+
```
|
|
60
|
+
/context-eng:verify-docs --fix
|
|
61
|
+
```
|
|
62
|
+
|
|
63
|
+
## Target Accuracy
|
|
64
|
+
|
|
65
|
+
| Level | Accuracy | Action |
|
|
66
|
+
|-------|----------|--------|
|
|
67
|
+
| Excellent | ≥80% | No action needed |
|
|
68
|
+
| Good | 60-80% | Update when convenient |
|
|
69
|
+
| Warning | 40-60% | Update soon |
|
|
70
|
+
| Critical | <40% | Update immediately |
|
|
71
|
+
|
|
72
|
+
## When to Run
|
|
73
|
+
|
|
74
|
+
- After refactoring
|
|
75
|
+
- After adding/removing significant code
|
|
76
|
+
- Before creating a PR
|
|
77
|
+
- Weekly maintenance
|