compound-agent 2.6.2 → 2.7.1

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/CHANGELOG.md CHANGED
@@ -7,6 +7,38 @@ All notable changes to this project will be documented in this file.
7
7
  The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
8
8
  and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
9
9
 
10
+ ## [2.7.1] - 2026-04-10
11
+
12
+ ### Added
13
+
14
+ - **Surface alignment reviewer**: New agent-role-skill that verifies cross-layer connectivity (frontend↔backend↔database↔API). Checks regenerate-and-diff compliance, architecture test presence, database testing fidelity, schema evolution safety, and auth surface coverage. Spawned for medium and large diffs during review phase.
15
+ - **P16 "Surfaces stay connected"**: New principle in the agentic audit/setup manifesto (Pillar II). Scores projects on cross-layer integration testing maturity (0-2). Updated overall scoring from 30→32.
16
+ - **5 deep research documents** shipped with the library: architecture tests (ArchUnit survey), regenerate-and-diff patterns, database testing patterns, test infrastructure as code, and protobuf schema evolution. Mapped to specific reviewer agents for calibrated reviews.
17
+ - **Research-calibrated review fleet**: Review skill now includes surface-alignment-reviewer in medium+ tier (9 reviewers), with calibration query for surface alignment lessons. Lesson-calibration references updated with all 5 research doc mappings.
18
+
19
+ ### Changed
20
+
21
+ - **build-great-things**: Added "System coherence is craft" as third design philosophy foundation. New structural coherence section in mandatory quality checklist (6 items). Added laziness pattern #13 "Disconnected layers".
22
+ - **Existing reviewers enhanced**: drift-detector, test-coverage-reviewer, and architecture-reviewer now reference relevant research documents. test-coverage-reviewer distinguishes integration from unit tests and flags SQLite substitution.
23
+
24
+ ### Fixed
25
+
26
+ - **Dependencies**: `modernc.org/sqlite` 1.48.0→1.48.1 (fixes memory leaks and double-free in multi-statement queries), `libc` 0.2.183→0.2.184 (patch bump of transitive Rust dependency).
27
+
28
+ ## [2.7.0] - 2026-04-09
29
+
30
+ ### Changed
31
+
32
+ - **Artifact consolidation**: All runtime artifacts (loop scripts, agent logs, phase state files) now live under `.compound-agent/` instead of being scattered across the project root and `.claude/`. The `ca setup` command auto-migrates legacy locations with conflict detection (skips if both old and new exist).
33
+ - **Root `.gitignore` management**: `ca setup` now maintains a marker-delimited block in the root `.gitignore` for `.compound-agent/` entries, replacing individual pattern lines.
34
+
35
+ ### Fixed
36
+
37
+ - **Retrocompatibility**: `GetPhaseState()` now falls back to the legacy `.claude/.ca-phase-state.json` path when the new `.compound-agent/` path is missing, and auto-migrates on first access. Repos upgrading without re-running `ca init` no longer lose active cook-it sessions.
38
+ - **Failure tracker on fresh repos**: `writeFailureState()` now auto-creates the `.compound-agent/` directory, fixing silent `post-tool-failure` hook regression on repos that haven't run `ca init`.
39
+ - **Stale path references**: Updated all `./infinity-loop.sh`, `./polish-loop.sh`, `./improvement-loop.sh` references to `.compound-agent/` prefix in GOTCHA.md, infinity-loop README, review-fleet.md, and generated script headers.
40
+ - **`.claude/.gitignore` cleanup**: Removed stale `.ca-*-state.json` patterns (files moved to `.compound-agent/`), added `.ca-hints-shown` and `skills/compound/skills_index.json`.
41
+
10
42
  ## [2.6.2] - 2026-04-05
11
43
 
12
44
  ### Added
package/README.md CHANGED
@@ -130,7 +130,7 @@ For systems too large for a single feature cycle. `/compound:architect` decompos
130
130
 
131
131
  # Step 2: generate and run the loop
132
132
  ca loop --reviewers claude-sonnet --review-every 3
133
- ./infinity-loop.sh
133
+ ./.compound-agent/infinity-loop.sh
134
134
  # → Processes each epic in dependency order: spec-dev → plan → work → review → compound
135
135
  # → Captures lessons after every cycle, improving subsequent cycles
136
136
  ```
@@ -154,7 +154,7 @@ ca loop --reviewers claude-sonnet --review-every 3
154
154
  ca loop --epics "beads-abc,beads-def,beads-ghi" --max-retries 2
155
155
 
156
156
  # Run it
157
- ./infinity-loop.sh
157
+ ./.compound-agent/infinity-loop.sh
158
158
  ```
159
159
 
160
160
  The loop respects beads dependency graphs — it only processes epics whose dependencies are complete. If an epic fails after `--max-retries` attempts, it stops and reports before proceeding.
@@ -180,10 +180,10 @@ ca improve --topics lint tests --max-iters 3
180
180
  ca improve --dry-run
181
181
 
182
182
  # Run the generated script
183
- ./improvement-loop.sh
183
+ ./.compound-agent/improvement-loop.sh
184
184
 
185
185
  # Preview without executing Claude sessions
186
- IMPROVE_DRY_RUN=1 ./improvement-loop.sh
186
+ IMPROVE_DRY_RUN=1 ./.compound-agent/improvement-loop.sh
187
187
  ```
188
188
 
189
189
  Each iteration makes one focused improvement, commits it, and moves on. If an iteration finds nothing to improve or fails validation, it reverts cleanly and moves to the next topic. The loop tracks consecutive no-improvement results and stops early to avoid diminishing returns.
@@ -303,7 +303,7 @@ The CLI binary is `ca` (alias: `compound-agent`).
303
303
  |---------|-------------|
304
304
  | `ca loop` | Generate infinity loop script for autonomous epic processing |
305
305
  | `ca loop --epics "id1,id2,id3"` | Target specific epic IDs (comma-separated) |
306
- | `ca loop -o <path>` | Custom output path (default: `./infinity-loop.sh`) |
306
+ | `ca loop -o <path>` | Custom output path (default: `./.compound-agent/infinity-loop.sh`) |
307
307
  | `ca loop --max-retries <n>` | Max retries per epic on failure (default: 1) |
308
308
  | `ca loop --force` | Overwrite existing script |
309
309
  | `ca loop --reviewers <names...>` | Enable review phase with specified reviewers (claude-sonnet, claude-opus, gemini, codex) |
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "compound-agent",
3
- "version": "2.6.2",
3
+ "version": "2.7.1",
4
4
  "type": "module",
5
5
  "description": "Learning system for Claude Code — avoids repeating mistakes across sessions",
6
6
  "bin": {
@@ -51,12 +51,12 @@
51
51
  "knowledge-management"
52
52
  ],
53
53
  "optionalDependencies": {
54
- "@syottos/darwin-arm64": "2.6.2",
55
- "@syottos/darwin-x64": "2.6.2",
56
- "@syottos/linux-arm64": "2.6.2",
57
- "@syottos/linux-x64": "2.6.2",
58
- "@syottos/win32-x64": "2.6.2",
59
- "@syottos/win32-arm64": "2.6.2"
54
+ "@syottos/darwin-arm64": "2.7.1",
55
+ "@syottos/darwin-x64": "2.7.1",
56
+ "@syottos/linux-arm64": "2.7.1",
57
+ "@syottos/linux-x64": "2.7.1",
58
+ "@syottos/win32-x64": "2.7.1",
59
+ "@syottos/win32-arm64": "2.7.1"
60
60
  },
61
61
  "author": "Nathan Delacrétaz",
62
62
  "license": "MIT",