@vheins/local-memory-mcp 0.14.2 → 0.14.4

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 (31) hide show
  1. package/dist/{chunk-T7WY5E4Y.js → chunk-U67YH7MG.js} +2 -2
  2. package/dist/dashboard/public/assets/{index-DHd85KEJ.js → index-CD8w6eaB.js} +27 -27
  3. package/dist/dashboard/public/index.html +1 -1
  4. package/dist/dashboard/server.js +1 -1
  5. package/dist/mcp/server.js +1 -1
  6. package/dist/prompts/architecture-design.md +22 -16
  7. package/dist/prompts/create-task.md +37 -0
  8. package/dist/prompts/csl-from-docs.md +9 -6
  9. package/dist/prompts/csl-scrapper.md +9 -6
  10. package/dist/prompts/documentation-sync.md +7 -0
  11. package/dist/prompts/export-task-to-github.md +20 -13
  12. package/dist/prompts/fix-suggestion.md +19 -17
  13. package/dist/prompts/import-github-issues.md +7 -0
  14. package/dist/prompts/learning-retrospective.md +5 -5
  15. package/dist/prompts/memory-agent-core.md +7 -0
  16. package/dist/prompts/memory-guided-review.md +6 -5
  17. package/dist/prompts/memory-index-policy.md +7 -0
  18. package/dist/prompts/project-briefing.md +7 -0
  19. package/dist/prompts/review-and-audit.md +14 -10
  20. package/dist/prompts/review-and-post-issue.md +13 -12
  21. package/dist/prompts/root-cause-analysis.md +12 -9
  22. package/dist/prompts/security-triage.md +12 -9
  23. package/dist/prompts/senior-code-review.md +9 -6
  24. package/dist/prompts/sentinel-issue-resolver.md +14 -5
  25. package/dist/prompts/session-planner.md +7 -6
  26. package/dist/prompts/task-management-guidelines.md +7 -0
  27. package/dist/prompts/task-memory-executor.md +52 -101
  28. package/dist/prompts/tech-affinity-scout.md +6 -5
  29. package/dist/prompts/technical-planning.md +45 -14
  30. package/dist/prompts/tool-usage-guidelines.md +7 -0
  31. package/package.json +4 -3
package/dist/dashboard/public/index.html CHANGED
@@ -8,7 +8,7 @@
8
8
  <link rel="preconnect" href="https://fonts.googleapis.com">
9
9
  <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
10
10
  <link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700;800&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">
11
- <script type="module" crossorigin src="/assets/index-DHd85KEJ.js"></script>
11
+ <script type="module" crossorigin src="/assets/index-CD8w6eaB.js"></script>
12
12
  <link rel="stylesheet" crossorigin href="/assets/index-CUg8rZCA.css">
13
13
  </head>
14
14
  <body>
package/dist/dashboard/server.js CHANGED
@@ -16,7 +16,7 @@ import {
16
16
  handleTaskClaim,
17
17
  listResources,
18
18
  logger
19
- } from "../chunk-T7WY5E4Y.js";
19
+ } from "../chunk-U67YH7MG.js";
20
20
 
21
21
  // src/dashboard/server.ts
22
22
  import express from "express";
package/dist/mcp/server.js CHANGED
@@ -60,7 +60,7 @@ import {
60
60
  toContextSlug,
61
61
  updateSessionFromInitialize,
62
62
  updateSessionRoots
63
- } from "../chunk-T7WY5E4Y.js";
63
+ } from "../chunk-U67YH7MG.js";
64
64
 
65
65
  // src/mcp/server.ts
66
66
  import readline from "readline";
package/dist/prompts/architecture-design.md CHANGED
@@ -3,25 +3,31 @@ name: architecture-design
3
3
  description: Plan system architecture, component layout, and data flow
4
4
  arguments:
5
5
  - name: tech_stack
6
- description: Technology stack. Optional — auto-detected from repo package files, language, and active task tags if omitted.
7
- required: false
6
+ description: Technology stack
7
+ required: true
8
8
  - name: requirements
9
- description: Key requirements. Optional — inferred from active task description, pending handoff, or recent conversation if omitted.
10
- required: false
9
+ description: Key requirements
10
+ required: true
11
11
  agent: System Architect
12
+ version: "1.0.0"
13
+ category: planning
14
+ tags: [architecture, system-design, components, data-flow, adr]
12
15
  ---
13
- ## 0. CONTEXT RESOLUTION
14
- - **tech_stack**: If provided, use directly. If omitted — detect from package.json, pyproject.toml, Gemfile, or repo file extensions.
15
- - **requirements**: If provided, use directly. If omitted — extract from active `in_progress` task description, pending handoff context, or recent conversation.
16
16
 
17
- Design system architecture for the active repository.
17
+ # Skill: Architecture Design
18
18
 
19
- Stack: (resolved above)
20
- Requirements: (resolved above)
19
+ > Design system architecture for repository.
21
20
 
22
- Output:
23
- 1. **Component Diagram**: Blocks & responsibilities.
24
- 2. **Data Flow**: Information movement.
25
- 3. **ADRs**: Rationale for patterns.
26
- 4. **Scalability/Reliability**: Growth & failure handling.
27
- 5. **Security**: Identity, protection, boundaries.
21
+ ## I/O
22
+ `tech_stack` (req), `requirements` (req) → `design/architecture/` (components, data flow, ADRs)
23
+
24
+ ## Rules
25
+ - Component Diagram: Blocks & responsibilities
26
+ - Data Flow: Information movement
27
+ - ADRs: Rationale for patterns
28
+ - Scalability/Reliability: Growth & failure handling
29
+ - Security: Identity, protection, boundaries
30
+
31
+ ## Chain
32
+ ← N/A
33
+ → `architecture-documentation`: `design/architecture/` → as-built architecture doc
package/dist/prompts/create-task.md CHANGED
@@ -6,6 +6,13 @@ arguments:
6
6
  description: Directive to analyze and break into tasks. Optional — derived from active task, pending handoff, or recent conversation if omitted.
7
7
  required: false
8
8
  agent: Task Planner
9
+ version: "1.0.0"
10
+ license: Proprietary — Personal Use Only
11
+ category: workflows
12
+ type: Orchestrator
13
+ complexity: Advanced
14
+ tags: [workflow, task-creation, planning, mcp]
15
+ author: vheins
9
16
  ---
10
17
 
11
18
  ## 🚫 FORBIDDEN: NON-EXECUTION
@@ -56,6 +63,36 @@ Log architectural/feature changes as `type: decision` via `memory-store`. Store
56
63
  - Parent/Child logic for complex directives.
57
64
  - Bulk limit: 500 records.
58
65
 
66
+ ## 5A. BLUEPRINT ORCHESTRATION IMPORT
67
+ When the instruction source is `idea-to-blueprint`:
68
+ - Create one root parent task for the full SDLC blueprint.
69
+ - Create phase parent tasks for P0 through P10 and final no-gap handoff.
70
+ - Create child tasks for every task breakdown row in the `idea-to-blueprint` skill.
71
+ - Recursively decompose every referenced skill until leaf-level tasks are reached.
72
+ - Preserve phase order with `depends_on`: P0 -> P1 -> G1 -> P2 -> G2 -> P3 -> G3 -> P4/P5 -> P6 -> P7 -> P8 -> P9 -> P10 -> GF.
73
+ - For conditional tasks, create them with explicit condition metadata instead of omitting them:
74
+ - UI design tasks run when the product has UI.
75
+ - accessibility tests run when UI exists.
76
+ - performance/load tests run when capacity risk exists.
77
+ - database migration tests run when schema changes exist.
78
+ - microservice/monolith tasks run only when architecture context applies.
79
+ - Gate tasks are internal quality checks. They should not ask for manual approval unless the gate is blocked by No-Go, missing mandatory input, or conflicting requirements.
80
+ - All task descriptions must use the strict Context & Analysis, Step & Implementation, Acceptance & Verification format.
81
+ - Call `task-list` before creation and link to existing related tasks instead of creating duplicates.
82
+ - Store a decision memory summarizing the blueprint task graph and major assumptions.
83
+
84
+ ## 5B. SPRINT PLAN IMPORT
85
+ When the instruction source is `.agents/documents/tasks/sprints/`:
86
+ - Treat sprint files and allocation audit as the source of truth.
87
+ - Create one root parent task for the full delivery scope.
88
+ - Create one sprint parent task per sprint and link each to the root via `parent_id`.
89
+ - Create module/feature child tasks under the sprint where the work is scheduled.
90
+ - Create atomic implementation/testing tasks under the matching module/feature parent.
91
+ - Convert sprint dependency columns and implementation order into `depends_on`.
92
+ - Preserve cross-sprint blockers by linking later sprint tasks to earlier sprint MCP task IDs.
93
+ - Call `task-list` before creation and skip duplicates; link to existing related tasks instead.
94
+ - Store a task creation audit in memory with created, linked, skipped duplicate, and blocked counts.
95
+
59
96
  ## 6. SELF-CHECK
60
97
  - ❌ No code/execution.
61
98
  - ✅ ONLY MCP tool calls.
package/dist/prompts/csl-from-docs.md CHANGED
@@ -3,15 +3,18 @@ name: csl-from-docs
3
3
  description: Create atomic CSL coding standards entries from a local file or directory path.
4
4
  arguments:
5
5
  - name: path
6
- description: Local path (file or directory) containing documentation or standards. Optional — defaults to docs/, README, or prompts definitions directory of the active repo if omitted.
7
- required: false
6
+ description: Local path (file or directory) containing documentation or standards.
7
+ required: true
8
8
  agent: Documentation Processor
9
+ version: "1.0.0"
10
+ category: workflows
11
+ tags: [workflow, csl, coding-standards, documentation, mcp]
9
12
  ---
10
- ## 0. CONTEXT RESOLUTION
11
- - **path**: If provided, use directly. If omitted — default to `docs/`, `README.md`, or `src/mcp/prompts/definitions/` in the active repo root.
12
- - **current_repo**: Auto-detect from git remote or active workspace context.
13
13
 
14
- Fetch and convert local documentation from the resolved path into atomic CSL (Coding Standards Library) entries.
14
+ Fetch and convert local documentation from the provided path into atomic CSL (Coding Standards Library) entries for the coding_standards entity.
15
+
16
+ Path: {{path}}
17
+ Current repo: {{current_repo}}
15
18
 
16
19
  Goal:
17
20
  - Analyze the provided path.
package/dist/prompts/csl-scrapper.md CHANGED
@@ -3,15 +3,18 @@ name: csl-scrapper
3
3
  description: Scrape trusted documentation from a URL into atomic CSL coding standards entries.
4
4
  arguments:
5
5
  - name: source_url
6
- description: Canonical URL for the documentation source to scrape. Optional — if omitted, use the most recently referenced URL from conversation context.
7
- required: false
6
+ description: Canonical URL for the documentation source to scrape.
7
+ required: true
8
8
  agent: Documentation Scraper
9
+ version: "1.0.0"
10
+ category: workflows
11
+ tags: [workflow, csl, scraping, coding-standards, mcp]
9
12
  ---
10
- ## 0. CONTEXT RESOLUTION
11
- - **source_url**: Use `{{source_url}}` if provided. If omitted — extract from the most recently mentioned URL in conversation context or active task description.
12
- - **current_repo**: Auto-detect from git remote or active workspace context.
13
13
 
14
- Fetch and convert trusted documentation from the resolved URL into atomic CSL (Coding Standards Library) entries.
14
+ Fetch and convert trusted documentation from the provided URL into atomic CSL (Coding Standards Library) entries for the coding_standards entity.
15
+
16
+ Source URL: {{source_url}}
17
+ Current repo: {{current_repo}}
15
18
 
16
19
  Goal:
17
20
  - Use the web_fetch tool (if available) to retrieve the content of the provided Source URL.
package/dist/prompts/documentation-sync.md CHANGED
@@ -3,6 +3,13 @@ name: documentation-sync
3
3
  description: Sync memory decisions with repository markdown files
4
4
  arguments: []
5
5
  agent: Documentation Specialist
6
+ version: "1.0.0"
7
+ license: Proprietary — Personal Use Only
8
+ category: workflows
9
+ type: Utility
10
+ complexity: Beginner
11
+ tags: [workflow, documentation, memory, sync]
12
+ author: vheins
6
13
  ---
7
14
  Reconcile local documentation with stored memories.
8
15
 
package/dist/prompts/export-task-to-github.md CHANGED
@@ -2,32 +2,39 @@
2
2
  name: export-task-to-github
3
3
  description: Export local tasks to GitHub Issues
4
4
  arguments:
5
+ - name: owner
6
+ description: GitHub repo owner
7
+ required: true
8
+ - name: repo
9
+ description: GitHub repo name
10
+ required: true
5
11
  - name: task_id
6
- description: Local task ID. Optional — if omitted, all tasks in the active repo are exported.
7
- required: false
12
+ description: Local task ID
13
+ required: true
8
14
  agent: Integration Architect
15
+ version: "1.0.0"
16
+ category: workflows
17
+ tags: [workflow, github, task-sync, mcp]
9
18
  ---
10
19
 
11
- ## 1. IDENTIFY ACTIVE PROJECT
12
- 1. **Detect**: Get repo name and owner from git remote (e.g. `git remote get-url origin`) or active workspace context. Parse `owner` and `repo` from the remote URL automatically.
13
- 2. **Verify**: Confirm the detected `owner`/`repo` before proceeding.
20
+ # Skill: export-task-to-github
14
21
 
15
- ## 2. RETRIEVE
16
- 1. **Scope**: If `task_id` provided — fetch that single task via `task-detail`. If omitted — call `task-list` (status: `pending,in_progress,completed`) to get all tasks in the active repo, then process each.
17
- 2. **Verify**: Ensure title/description exist per task. Use `memory-search` for gaps.
22
+ ## 1. RETRIEVE
23
+ 1. **Fetch**: Call `task-detail` for `task_id`.
24
+ 2. **Verify**: Ensure title/description exist. Use `memory-search` for gaps.
18
25
 
19
- ## 3. SYNC CHECK
20
- 1. **Search**: Use `search_issues` for `task_code` scoped to detected `owner`/`repo`.
26
+ ## 2. SYNC CHECK
27
+ 1. **Search**: Use `search_issues` for `task_code`.
21
28
  2. **De-duplicate**: If issue exists, update local task `metadata` with URL. DO NOT re-create.
22
29
 
23
- ## 4. CREATE ISSUE
30
+ ## 3. CREATE ISSUE
24
31
  If new:
25
- - **Write**: Use `issue_write` (method: 'create') with detected `owner`/`repo`.
32
+ - **Write**: Use `issue_write` (method: 'create').
26
33
  - **Content**: Match local title/body exactly.
27
34
  - **Traceability**: Append `task_code` and `task_id` to body.
28
35
  - **Comments**: Post local comments via `add_issue_comment`.
29
36
 
30
- ## 5. LINK
37
+ ## 4. LINK
31
38
  - **Update**: Call `task-update`.
32
39
  - **Metadata**: Add GitHub URL.
33
40
  - **Comment**: "Exported to GitHub Issue #X".
package/dist/prompts/fix-suggestion.md CHANGED
@@ -3,27 +3,29 @@ name: fix-suggestion
3
3
  description: Targeted fix with before/after code and test case.
4
4
  arguments:
5
5
  - name: tech_stack
6
- description: Target tech stack. Optional — inferred from repo/context if omitted.
7
- required: false
6
+ description: Target tech stack.
7
+ required: true
8
8
  - name: bug_description
9
- description: Bug behavior. Optional — inferred from active conversation or task context if omitted.
10
- required: false
9
+ description: Bug behavior.
10
+ required: true
11
11
  - name: root_cause
12
- description: Identified root cause. Optional — inferred from recent error/log context if omitted.
13
- required: false
12
+ description: Identified root cause.
13
+ required: true
14
14
  agent: Debugging Expert
15
+ version: "1.1.0"
16
+ category: debugging
17
+ tags: [fix, patch, bug-fix, code-change, test-case, debugging]
15
18
  ---
16
19
 
17
- ## 0. CONTEXT RESOLUTION
18
- Resolve missing arguments from available context before proceeding:
19
- - **tech_stack**: Detect from repo language, package files, or active task tags. Fallback: ask agent to infer from open files.
20
- - **bug_description**: Extract from active task description, recent conversation, or error logs. Fallback: describe observable broken behavior.
21
- - **root_cause**: Extract from recent analysis, error traces, or `memory-search` results. Fallback: state "unknown — investigation required".
20
+ # Skill: Fix Suggestion
22
21
 
23
- Provide precise, minimal fix for confirmed bug.
22
+ > Provide precise, minimal fix for confirmed bug.
24
23
 
25
- Output:
26
- 1. **Explanation**: Why it happens & how fix works.
27
- 2. **Before/After**: Diff style code blocks with comments.
28
- 3. **Checklist**: Meta changes (config, migrations).
29
- 4. **Verification**: Regression test case.
24
+ ## I/O
25
+ tech_stack (req), bug_description (req), root_cause (req) → diagnosis report + fix suggestion
26
+
27
+ ## Rules
28
+ 1. Explanation: Why it happens and how fix works
29
+ 2. Before/After: Diff style code blocks with comments
30
+ 3. Checklist: Meta changes (config, migrations)
31
+ 4. Verification: Regression test case
package/dist/prompts/import-github-issues.md CHANGED
@@ -3,6 +3,13 @@ name: import-github-issues
3
3
  description: Import GitHub Issues as local tasks.
4
4
  arguments: []
5
5
  agent: Integration Scout
6
+ version: "1.0.0"
7
+ license: Proprietary — Personal Use Only
8
+ category: workflows
9
+ type: Orchestrator
10
+ complexity: Intermediate
11
+ tags: [workflow, github, issue-import, mcp]
12
+ author: vheins
6
13
  ---
7
14
 
8
15
  ## 1. FETCH
package/dist/prompts/learning-retrospective.md CHANGED
@@ -3,15 +3,15 @@ name: learning-retrospective
3
3
  description: Harvest knowledge from completed work.
4
4
  arguments:
5
5
  - name: task_id
6
- description: ID or code of completed task. Optional — defaults to most recently completed task in the active repo.
6
+ description: ID of completed task.
7
7
  required: false
8
8
  agent: Knowledge Harvester
9
+ version: "1.0.0"
10
+ category: workflows
11
+ tags: [workflow, retrospective, memory, knowledge-management]
9
12
  ---
10
- ## 0. CONTEXT RESOLUTION
11
- 1. **Repo**: Auto-detect from git remote or active workspace context. All MCP calls MUST be scoped to this repo.
12
- 2. **Task**: If `task_id` provided — use it directly. If omitted — call `task-list` (status: `completed`, limit: 1, ordered by updated_at desc) to get the most recently completed task.
13
13
 
14
- Extract durable knowledge from the resolved task for the active repository.
14
+ Extract durable knowledge from task {{task_id}} for repository.
15
15
 
16
16
  Identify and `memory-store`:
17
17
  1. **Mistakes**: Hard-to-find bugs or environment quirks.
package/dist/prompts/memory-agent-core.md CHANGED
@@ -3,6 +3,13 @@ name: memory-agent-core
3
3
  description: Behavioral contract for memory-aware agents.
4
4
  arguments: []
5
5
  agent: Memory Guardian
6
+ version: "1.0.0"
7
+ license: Proprietary — Personal Use Only
8
+ category: workflows
9
+ type: Utility
10
+ complexity: Advanced
11
+ tags: [memory, workflow, guardrails, mcp]
12
+ author: vheins
6
13
  ---
7
14
  # Memory Guardian Protocol
8
15
 
package/dist/prompts/memory-guided-review.md CHANGED
@@ -3,14 +3,15 @@ name: memory-guided-review
3
3
  description: Review code for compliance with stored decisions.
4
4
  arguments:
5
5
  - name: file_path
6
- description: File to review. Optional — if omitted, review the currently open/active file in the workspace, or all recently modified files.
7
- required: false
6
+ description: File to review.
7
+ required: true
8
8
  agent: Code Auditor
9
+ version: "1.0.0"
10
+ category: coding
11
+ tags: [code-review, memory, compliance, mcp]
9
12
  ---
10
- ## 0. CONTEXT RESOLUTION
11
- - **file_path**: If provided, use it. If omitted — use the currently active/open file from workspace context, or list recently modified files via git and process them.
12
13
 
13
- Audit the resolved file(s) against stored project knowledge.
14
+ Audit {{file_path}} against stored project knowledge.
14
15
 
15
16
  Steps:
16
17
  1. **Search Memory**: Call `memory-search` using `current_file_path='{{file_path}}'`.
package/dist/prompts/memory-index-policy.md CHANGED
@@ -3,6 +3,13 @@ name: memory-index-policy
3
3
  description: Strict memory storage criteria.
4
4
  arguments: []
5
5
  agent: Memory Auditor
6
+ version: "1.0.0"
7
+ license: Proprietary — Personal Use Only
8
+ category: workflows
9
+ type: Utility
10
+ complexity: Beginner
11
+ tags: [memory, indexing, policy, mcp]
12
+ author: vheins
6
13
  ---
7
14
  # Memory Indexing Rules
8
15
 
package/dist/prompts/project-briefing.md CHANGED
@@ -3,6 +3,13 @@ name: project-briefing
3
3
  description: Contextual onboarding to current repository.
4
4
  arguments: []
5
5
  agent: Session Concierge
6
+ version: "1.0.0"
7
+ license: Proprietary — Personal Use Only
8
+ category: workflows
9
+ type: Utility
10
+ complexity: Beginner
11
+ tags: [workflow, briefing, onboarding, memory, backlog]
12
+ author: vheins
6
13
  ---
7
14
  Initialize session in repository.
8
15
 
package/dist/prompts/review-and-audit.md CHANGED
@@ -3,20 +3,24 @@ name: review-and-audit
3
3
  description: Audit documentation against implementation; generate local tasks for gaps.
4
4
  arguments:
5
5
  - name: target
6
- description: Module, feature, or component to audit. Optional — if omitted, audits all available documentation against the full implementation.
6
+ description: Module, feature, or component to audit.
7
7
  required: false
8
8
  agent: Quality Auditor
9
+ version: "1.0.0"
10
+ license: Proprietary — Personal Use Only
11
+ category: workflows
12
+ type: Orchestrator
13
+ complexity: Advanced
14
+ tags: [workflow, audit, ux, gap-analysis, mcp]
15
+ author: vheins
9
16
  ---
10
-
11
- ## 0. CONTEXT RESOLUTION
12
- - **Target**: If `target` provided — scope audit to that module/feature/component. If omitted — **fallback**: audit ALL existing documentation (README, docs/, prompts, schemas) against the full codebase implementation. Enumerate each doc file and compare with corresponding source.
13
- - **Repo**: Auto-detect from git remote or active workspace context.
17
+ # Skill: review-and-audit (Audit Agent)
14
18
 
15
19
  ## 1. ANALYSIS
16
- 1. **Discovery**: Explore docs and code. Coding sub-agents MAY be used for parallel file reading if the agent supports sub-agents. `chrome-devtools-mcp` MCP tools (direct tool calls) are **ALLOWED**. **FORBIDDEN**: spawning a `browser_subagent` — do NOT invoke the `browser_subagent` tool here. Distinction: MCP tool call = allowed; spawned browser agent process = forbidden.
17
- 2. **UX Audit**: If target involves UI and the audit explicitly requires visual inspection, note it as a separate task — do NOT block the audit on it.
20
+ 1. **Sequential Discovery**: Explore docs and code sequentially. NO parallel sub-agents.
21
+ 2. **UX Audit**: Use `chrome-dev-tools` for visual, navigation, and responsiveness checks.
18
22
  3. **Reference Audit**: Check current tool/prompt/resource definitions through code or the dashboard Reference flow.
19
- 4. **Compare**: Match docs + code findings to find gaps/misalignments.
23
+ 4. **Compare**: Match docs + code findings against live UI to find gaps/misalignments.
20
24
 
21
25
  ## 🚫 FORBIDDEN: NON-EXECUTION
22
26
  DO NOT edit/create/delete files, run commands, or implement code.
@@ -40,8 +44,8 @@ ONLY call MCP tools. No prose, code, or external plans.
40
44
  - **Finding**: Gap trigger.
41
45
  - **Observation**: Reasoning.
42
46
  - **Goal**: Clear objective.
43
- ### 2. Step & Implementation
44
- - Detailed execution steps per path/layer.
47
+ ### 2. Target Files & Implementation
48
+ - Combined scope/steps per path/layer.
45
49
  ### 3. Acceptance & Verification
46
50
  - **Checklist**: `[ ]` criteria.
47
51
  - **Testing**: Scenarios.
package/dist/prompts/review-and-post-issue.md CHANGED
@@ -3,25 +3,26 @@ name: review-and-post-issue
3
3
  description: Audit documentation against implementation; generate GitHub issues for gaps.
4
4
  arguments:
5
5
  - name: owner
6
- description: GitHub repo owner. Optional — auto-detected from git remote if omitted.
7
- required: false
6
+ description: GitHub repo owner.
7
+ required: true
8
8
  - name: repo
9
- description: GitHub repo name. Optional — auto-detected from git remote if omitted.
10
- required: false
9
+ description: GitHub repo name.
10
+ required: true
11
11
  - name: target
12
- description: Module, feature, or component to audit. Optional — if omitted, audits all documentation against the full implementation.
12
+ description: Module, feature, or component to audit.
13
13
  required: false
14
14
  agent: Quality Auditor
15
+ version: "1.0.0"
16
+ category: workflows
17
+ tags: [workflow, audit, github, issue-triage]
15
18
  ---
16
19
 
17
- ## 0. CONTEXT RESOLUTION
18
- 1. **Owner/Repo**: Auto-detect from `git remote get-url origin` or active workspace context. Parse `owner` and `repo` from the remote URL. Verify before proceeding.
19
- 2. **Target**: If `target` provided — scope to that module/feature. If omitted — **fallback**: audit ALL existing documentation against the full codebase.
20
+ # Skill: review-and-post-issue (Audit Agent)
20
21
 
21
22
  ## 1. ANALYSIS
22
- 1. **Discovery**: Explore docs and code. Coding sub-agents MAY be used for parallel reading if the agent supports sub-agents. **FORBIDDEN: browser sub-agents**.
23
- 2. **UX Audit**: If the target involves UI, note it as a separate GitHub issue — do NOT block audit on visual inspection.
24
- 3. **Compare**: Match docs + code findings to identify gaps/misalignments.
23
+ 1. **Sequential Discovery**: Explore docs and code sequentially. NO parallel sub-agents.
24
+ 2. **UX Audit**: If applicable, use `chrome-dev-tools` for visual, navigation, and responsiveness checks.
25
+ 3. **Compare**: Match findings against live UI to find gaps/misalignments.
25
26
 
26
27
  ## 🚫 FORBIDDEN: NON-EXECUTION
27
28
  DO NOT edit/create/delete files, run commands, or implement code.
@@ -42,7 +43,7 @@ No prose. No external plans.
42
43
  - **Finding**: Gap trigger.
43
44
  - **Observation**: Reasoning.
44
45
  - **Goal**: Clear objective.
45
- ### 2. Step & Implementation
46
+ ### 2. Target Files & Implementation
46
47
  - Path/layer specific changes.
47
48
  ### 3. Acceptance & Verification
48
49
  - **Checklist**: `[ ]` criteria.
package/dist/prompts/root-cause-analysis.md CHANGED
@@ -3,22 +3,25 @@ name: root-cause-analysis
3
3
  description: 5-Why analysis to trace bug origins.
4
4
  arguments:
5
5
  - name: tech_stack
6
- description: Target tech stack. Optional — auto-detected from repo/context if omitted.
7
- required: false
6
+ description: Target tech stack.
7
+ required: true
8
8
  - name: bug_description
9
- description: Bug behavior. Optional — inferred from active task or recent conversation if omitted.
10
- required: false
9
+ description: Bug behavior.
10
+ required: true
11
11
  - name: symptoms
12
12
  description: Logs, errors, metrics.
13
13
  required: false
14
14
  agent: Diagnostic Lead
15
+ version: "1.0.0"
16
+ category: debugging
17
+ tags: [root-cause, 5-why, debugging, diagnosis]
15
18
  ---
16
- ## 0. CONTEXT RESOLUTION
17
- - **tech_stack**: If provided, use directly. If omitted — detect from repo package files or active task tags.
18
- - **bug_description**: If provided, use directly. If omitted — extract from active task description or recent error/log context.
19
- - **symptoms**: Optional. Use if provided; otherwise infer from available logs or error traces.
20
19
 
21
- Conduct root cause analysis for the active repository bug.
20
+ Conduct root cause analysis for repository bug.
21
+
22
+ Stack: {{tech_stack}}
23
+ Bug: {{bug_description}}
24
+ Symptoms: {{symptoms}}
22
25
 
23
26
  Output:
24
27
  1. **Symptom**: Technical problem restatement.
package/dist/prompts/security-triage.md CHANGED
@@ -3,22 +3,25 @@ name: security-triage
3
3
  description: Assess vulnerability exploitability and prioritize fix.
4
4
  arguments:
5
5
  - name: tech_stack
6
- description: App stack. Optional — auto-detected from repo/context if omitted.
7
- required: false
6
+ description: App stack.
7
+ required: true
8
8
  - name: vulnerability_report
9
- description: Report details (CVE, SAST). Optional — extracted from active task description or recent conversation if omitted.
10
- required: false
9
+ description: Report details (CVE, SAST).
10
+ required: true
11
11
  - name: codebase_context
12
12
  description: Usage context.
13
13
  required: false
14
14
  agent: Security Engineer
15
+ version: "1.0.0"
16
+ category: debugging
17
+ tags: [security, triage, vulnerability, cvss, appsec]
15
18
  ---
16
- ## 0. CONTEXT RESOLUTION
17
- - **tech_stack**: If provided, use directly. If omitted — detect from repo package files, language, or active task tags.
18
- - **vulnerability_report**: If provided, use directly. If omitted — extract from active task description, recent conversation, or attached SAST output.
19
- - **codebase_context**: Optional. Use if provided.
20
19
 
21
- Triage the resolved vulnerability for the active repository.
20
+ Triage vulnerability for repository.
21
+
22
+ Stack: {{tech_stack}}
23
+ Report: {{vulnerability_report}}
24
+ Context: {{codebase_context}}
22
25
 
23
26
  Output:
24
27
  1. **Classification**: Type, CVE, CVSS, vector.
package/dist/prompts/senior-code-review.md CHANGED
@@ -3,18 +3,21 @@ name: senior-code-review
3
3
  description: Comprehensive production-readiness evaluation.
4
4
  arguments:
5
5
  - name: tech_stack
6
- description: Tech stack. Optional — auto-detected from repo package files, language, or active task tags if omitted.
7
- required: false
6
+ description: Tech stack.
7
+ required: true
8
8
  - name: context
9
9
  description: Production context (SLA, data, conventions).
10
10
  required: false
11
11
  agent: Principal Reviewer
12
+ version: "1.1.0"
13
+ category: coding
14
+ tags: [code-review, production-readiness, security, observability, senior-review, architecture]
12
15
  ---
13
- ## 0. CONTEXT RESOLUTION
14
- - **tech_stack**: If provided, use directly. If omitted — detect from repo package files or language.
15
- - **context**: Optional. Use if provided; otherwise infer from task description or recent conversation.
16
16
 
17
- Perform production-readiness review for the active repository.
17
+ Perform production-readiness review for repository.
18
+
19
+ Stack: {{tech_stack}}
20
+ Context: {{context}}
18
21
 
19
22
  Audit Dimensions:
20
23
  1. **Errors**: Completeness & patterns.
package/dist/prompts/sentinel-issue-resolver.md CHANGED
@@ -6,6 +6,13 @@ arguments:
6
6
  description: The full URL of the GitHub issue to resolve.
7
7
  required: true
8
8
  agent: SENTINEL Issue Resolver
9
+ version: "1.0.0"
10
+ license: Proprietary — Personal Use Only
11
+ category: workflows
12
+ type: Orchestrator
13
+ complexity: Advanced
14
+ tags: [workflow, github, issue-resolution, sentinel]
15
+ author: vheins
9
16
  ---
10
17
 
11
18
  # SENTINEL Protocol
@@ -26,12 +33,14 @@ You are **SENTINEL**, an elite issue resolution agent. Your primary objective is
26
33
  1. **Identity**: Use the local Git configuration (name/email) for all commits.
27
34
  2. **Commit Format**: Every commit MUST follow this specific structure:
28
35
  ```
29
- type(scope): [task-code] fix #{{issue_number}} - your commit message
30
-
31
- - [Task Title]
32
- [Summary Task]
36
+ type(scope): your commit message
37
+
38
+ - {{task_title}}
39
+ {{summary_task}}
40
+
41
+ {{keyword}} #{{issue_number}}
33
42
  ```
34
- *Note: Extract the issue number from the provided `issue_url`.*
43
+ Use `fix` for bug fixes, `closes` for features/chores, `resolve` as general. Extract the issue number from the provided `issue_url`.
35
44
 
36
45
  3. **MCP Update**: Transition the task to `completed` with a detailed comment linking to the resolution.
37
46
  4. **Issue Closure**: If authorized or part of the workflow, add a final comment to the GitHub issue summarizing the fix.