@namch/agent-assistant 1.0.3 → 1.0.4

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 (5) hide show
  1. package/CHANGELOG.md +54 -0
  2. package/commands/code/focus.md +20 -0
  3. package/commands/code/hard.md +20 -0
  4. package/commands/code.md +3 -0
  5. package/package.json +4 -2
package/CHANGELOG.md ADDED
@@ -0,0 +1,54 @@
1
+ # Changelog
2
+
3
+ All notable changes to this project will be documented in this file.
4
+
5
+ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
6
+ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
7
+
8
+ ## [1.0.4] - 2026-01-30
9
+
10
+ ### Added
11
+
12
+ - **Plan-already-provided short-circuit**: `/code:hard` and `/code:focus` now detect when the user references an existing plan (`@plan`, `@PLAN-...`, path to `PLAN-*.md`, or phrases like "according to plan" / "follow the plan"). When a valid plan exists, **research, scout, and brainstorm phases are skipped** and execution goes straight to context optimization → implementation → test → review.
13
+
14
+ ### Changed
15
+
16
+ - **`commands/code.md`**: Routing logic updated so that when the user references an existing plan, the router directs to `/code:hard` or `/code:focus` (workflow then skips redundant phases).
17
+ - **`commands/code/hard.md`** and **`commands/code/focus.md`**: New section "PLAN-ALREADY-PROVIDED: SKIP REDUNDANT PHASES" with detection rules and resolution (skip Phase 1–3 when plan provided).
18
+
19
+ ## [1.0.3] - 2026-01-30
20
+
21
+ ### Added
22
+
23
+ - **Reporter Agent**: New `reporter` agent for documentation and reporting (create/update reports, template-based output).
24
+ - **Report Command**: `/report` with variants `auto`, `fast`, `hard`, `focus` — status updates, deep analysis, and focus mode with context optimization.
25
+ - **Focus Variant**: For all applicable commands (`cook`, `code`, `fix`, `debug`, `design`, `plan`, `test`) — **Clear context** and **auto-run phases** for guaranteed clean execution without context rot.
26
+ - **Matrix-Skills**: Updated `matrix-skills/_index.yaml` with reporter profile and domain mappings.
27
+
28
+ ### Changed
29
+
30
+ - **Commands**: Router workflows now support **Clear context** and **focus** variant (force clear context, auto run phase) across cook, code, fix, debug, design, plan, test.
31
+ - **Documentation**: README, AGENT.md, CLAUDE.md, CURSOR.md, COPILOT.md, GEMINI.md, rules, code-assistants, web data, and documents updated for reporter, `/report`, and focus variants.
32
+
33
+ ## [1.0.2] - 2026-01-30
34
+
35
+ ### Added
36
+
37
+ - **Web Integration**: Added `web` directory content and resources.
38
+ - **Documentation**: Updated `README.md` with comprehensive usage instructions and project details.
39
+
40
+ ## [1.0.1] - 2026-01-29
41
+
42
+ ### Added
43
+
44
+ - **Matrix Skills Integration**: Implemented a massive library of 2000+ specialized skills (`matrix-skills`) to enhance agent capabilities.
45
+ - Improved skill discovery and routing mechanisms.
46
+
47
+ ## [1.0.0] - 2026-01-26
48
+
49
+ ### Added
50
+
51
+ - **Initial Release**: First stable release of `@namch/agent-assistant`.
52
+ - **Core Orchestration**: Framework for managing multi-agent workflows.
53
+ - **CLI Tool**: `agent-assistant` CLI for easy installation and management.
54
+ - **Multi-Assistant Support**: Compatibility with Cursor, GitHub Copilot, and Claude Code (Antigravity).
package/commands/code/focus.md CHANGED
@@ -28,6 +28,26 @@ execution-mode: execute
28
28
 
29
29
  ---
30
30
 
31
+ ## 📌 PLAN-ALREADY-PROVIDED: SKIP REDUNDANT PHASES
32
+
33
+ **Before starting Phase 1**, resolve whether the user is **coding from an existing plan**. If yes, **do not** run research, scout, or brainstorm.
34
+
35
+ ### Detection
36
+
37
+ - User message references a plan: e.g. `@plan`, `@PLAN-...`, path like `./reports/plans/PLAN-{name}.md`, or phrases like "according to plan", "follow the plan", "code from plan", "implement per plan".
38
+ - Or a plan file already exists for this task at `./reports/plans/PLAN-{task}.md` (derive `{task}` from `$ARGUMENTS` or from the referenced file name).
39
+
40
+ ### Resolution
41
+
42
+ 1. **CHECK**: Does a valid plan file exist (user-provided path or `./reports/plans/PLAN-{task}.md`)?
43
+ 2. **IF YES**:
44
+ - **SKIP** Phase 1 (Requirements/Brainstorm), Phase 2 (Scout), Phase 3 (Planning).
45
+ - **ANNOUNCE**: "✅ Plan provided/found — skipping research, scout, and planning. Proceeding to implementation."
46
+ - **GO TO**: Strict Context Gate (automatic) → then Phase 4 (Implementation) → Phase 5 (Testing) → Phase 6 (Review).
47
+ 3. **IF NO**: Run phases in order: Phase 1 → 2 → 3 → Context Gate → 4 → 5 → 6.
48
+
49
+ ---
50
+
31
51
  ## 🔀 TIERED EXECUTION PROTOCOL (MANDATORY)
32
52
 
33
53
  > **Reference**: `{RULES_PATH}/ADAPTIVE-EXECUTION.md`
package/commands/code/hard.md CHANGED
@@ -25,6 +25,26 @@ execution-mode: execute
25
25
 
26
26
  ---
27
27
 
28
+ ## 📌 PLAN-ALREADY-PROVIDED: SKIP REDUNDANT PHASES
29
+
30
+ **Before starting Phase 1**, resolve whether the user is **coding from an existing plan**. If yes, **do not** run research, scout, or brainstorm.
31
+
32
+ ### Detection
33
+
34
+ - User message references a plan: e.g. `@plan`, `@PLAN-...`, path like `./reports/plans/PLAN-{name}.md`, or phrases like "according to plan", "follow the plan", "code from plan", "implement per plan".
35
+ - Or a plan file already exists for this task at `./reports/plans/PLAN-{task}.md` (derive `{task}` from `$ARGUMENTS` or from the referenced file name).
36
+
37
+ ### Resolution
38
+
39
+ 1. **CHECK**: Does a valid plan file exist (user-provided path or `./reports/plans/PLAN-{task}.md`)?
40
+ 2. **IF YES**:
41
+ - **SKIP** Phase 1 (Requirements/Brainstorm), Phase 2 (Scout), Phase 3 (Planning).
42
+ - **ANNOUNCE**: "✅ Plan provided/found — skipping research, scout, and planning. Proceeding to implementation."
43
+ - **GO TO**: Verification Checkpoint (Context Optimization) → then Phase 4 (Implementation) → Phase 5 (Testing) → Phase 6 (Review).
44
+ 3. **IF NO**: Run phases in order: Phase 1 → 2 → 3 → Checkpoint → 4 → 5 → 6.
45
+
46
+ ---
47
+
28
48
  ## 🔀 TIERED EXECUTION PROTOCOL (MANDATORY)
29
49
 
30
50
  > **Reference**: `{RULES_PATH}/ADAPTIVE-EXECUTION.md`
package/commands/code.md CHANGED
@@ -28,6 +28,9 @@ execution-mode: router
28
28
  ## ROUTING LOGIC
29
29
 
30
30
  ```
31
+ IF user references an existing plan (@plan, @PLAN-..., path to PLAN-*.md, or "according to plan" / "follow the plan"):
32
+ → Route to /code:hard or /code:focus (workflow will SKIP research, scout, brainstorm and go straight to implementation)
33
+
31
34
  IF task is simple (clear requirements, single file):
32
35
  → Route to /code:fast
33
36
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@namch/agent-assistant",
3
- "version": "1.0.3",
3
+ "version": "1.0.4",
4
4
  "description": "Multi-agent orchestration framework for AI coding assistants (Cursor, Copilot, Antigravity, Claude Code).",
5
5
  "main": "cli/install.js",
6
6
  "bin": {
@@ -44,7 +44,9 @@
44
44
  "skills/",
45
45
  "documents/",
46
46
  "code-assistants/",
47
- "LICENSE"
47
+ "LICENSE",
48
+ "README.md",
49
+ "CHANGELOG.md"
48
50
  ],
49
51
  "devDependencies": {
50
52
  "@semantic-release/changelog": "^6.0.3",