7f5b5d8eca
v6.2.0: agents rewritten for CC 2.1.269 + Claude 5 prompting; PQ pass; memory-sync prompting-guide
2026-09-12 20:22:56 +01:00
421113f661
fix(agent-router): evidence-gated denies, weak signals, override escape hatch
2026-08-16 21:31:16 +01:00
eab5cf443a
v6.1.3: manager-setup - gate HARD wall arm behind explicit confirm
2026-08-16 20:45:36 +01:00
bdd30823da
feat(setup-status): group run-list into 6 dependency-aware stages
2026-08-16 19:22:18 +01:00
8a06abe880
v6.1.1: fail-closed MCP classification in HARD-wall guard + mcpAllow recovery key
2026-08-16 18:18:25 +01:00
bfb24ca0ab
v6.1.0: brewtools:context-slim - permanent context compression
2026-08-16 17:46:21 +01:00
765ccc1e7c
v6.0.0: audit re-verification, 4 fix waves, agent compaction, full docs pass
2026-08-16 15:18:22 +01:00
225e1ad63f
v5.7.0: post-compaction role/plan/task recall hooks + agent-router content_version staleness
2026-08-15 09:26:24 +01:00
8bb41ebe4b
v5.6.1: think-short subagent injection moved to SubagentStart additionalContext
2026-08-14 19:04:42 +01:00
34a278290a
feat(brewcode,brewdoc,brewtools): automatic content_version staleness detection for all setup skills
2026-08-13 08:16:21 +01:00
c3619ec19c
v5.5.3: hard wall no longer bypassed by claude --agent sessions
2026-08-11 21:36:02 +01:00
6d2e357b01
v5.5.2: semble-first rule compressed ~3x, hard size budget enforced by test J9
2026-08-10 18:36:46 +01:00
267f566503
fix(skills): generated skills stay model-invocable
2026-08-10 18:21:03 +01:00
20937c5a83
v5.5.0: agent-return-setup - subagent return budget hooks, Return Contract on all 8 shipped agents, docs wave
2026-08-10 17:43:39 +01:00
60cc11cdc1
v5.4.0: prompt-first contract for every skill - free-form prompt in position 1, deterministic mode resolution, mandatory PLAN block, validator checks 7-10
2026-08-10 11:40:46 +01:00
957d8ab61c
v5.3.2: skill frontmatter names go bare - double plugin namespace in the picker fixed, validator enforces name==dir
2026-08-10 08:41:51 +01:00
cee852ca4d
v5.3.1: agentOf reads agent_id in both hooks, sub records stamp aid, pre-fix telemetry quarantined
2026-08-09 18:35:55 +01:00
58f37d2fd0
v5.3.0: semble reminder retuned to a high-specificity gate, delivery inside subagents proven, two hypotheses refuted
...
Reminder default reminderEvery 5 -> 1, injection 65 -> 29 tokens, gate replay
on 2543 calls: 14/14 (was 1024/204). Installer CLAUDE.md conflict scan added,
but its motivating hypothesis failed A/B (p=0.45, n=32). Reminder fired 0/59
in 7 sessions - conversion undefined. SubagentStart is the channel that works
(8/8 first-tool). Docs across mdx/README/SKILL/INSTALL reconciled.
2026-08-09 17:53:33 +01:00
fa32a58f6e
v5.2.4: reconcile three drifted hook helpers - pruneStale symlink hardening, safeSegment numeric ids, recordTouched dedupe
2026-08-09 16:14:43 +01:00
f313d71b78
fix(semble): mark PostToolUseFailure open/search records with fail:true
2026-08-09 14:54:52 +01:00
9ddf376e77
v5.2.2: placeholder version stamps hard-fail, dead brewdoc logger removed, brewdoc README enable/disable claim corrected
2026-08-09 14:41:40 +01:00
5f7746a0f8
fix(semble): derive cache content set from SEMBLE_CONTENT_ARGS
2026-08-09 14:40:42 +01:00
3abdf6bfe9
v5.2.0: semble advisory hooks restored - 6 settings entries, every-Nth cadence, delivery proven end-to-end
2026-08-09 10:55:33 +01:00
9ca3c628dd
v5.1.0: unified artifact metadata + version-based staleness across all 10 setups
2026-08-09 08:18:53 +01:00
3e4081107b
v5.0.0: -setup suffix on all setup skills, guide deleted, setup-status added, canonical mode vocabulary, critical-fix wave
2026-08-08 19:11:07 +01:00
30ae5af514
v4.10.0: delete brewcode:spec, superreview upgrade + self-sync, task-board progress doc, etalon-first injection
2026-08-08 15:52:18 +01:00
822e822718
refactor(brewcode): remove grepai skill, agent, hooks and all references
2026-08-02 17:51:22 +01:00
ea944a32fe
v4.1.0: add /brewdoc:docsync doc-staleness tracker, remove auto-sync
...
- New /brewdoc:docsync skill: user-run generator installs project-local hooks
(track/watch/gate) + config; tracks stale docs by last_updated date; modes
init/status/sync/reread/frontmatter/uninstall; confirm-before-sync gate
- Removed auto-sync skill, bd-auto-sync-processor agent, auto-sync:* frontmatter
tags (~44 files), and skill-creator template injection
- Docs updated at all levels (README, MDX, navigation, guide catalog, CLAUDE.md)
2026-07-19 13:16:50 +01:00
3055aaf33b
v4.0.0: strip injection hooks, native ${CLAUDE_PLUGIN_ROOT}, grepai self-install; remove dead mode system + brewtools debate/skill-toggle/agent-toggle; add superreview
2026-06-30 20:04:35 +01:00
a6c1b14965
v3.19.5: remove dead lock/task code from hooks; clean all stale refs
2026-06-29 20:19:30 +01:00
f6f93ea30e
v3.19.4: remove knowledge system + report templates + stale hook refs
2026-06-29 20:09:24 +01:00
25b852b346
v3.19.3: purge all plan/start refs from setup, agents, docs
2026-06-29 19:49:38 +01:00
955762c9c8
v3.19.1: restore skills/agents/hooks/templates stripped in v3.19.0 -- only plan+start stay removed (11 skills)
2026-06-29 19:25:28 +01:00
0a77b5bc3c
v3.19.0: strip task execution system -- keep spec+grepai only
...
- remove orphaned skills/agents skill (README commits to two skills)
- spec skill: drop /brewcode:setup + /brewcode:plan deps, hand SPEC to developer agent
- grepai skill/docs: drop /brewcode:setup prereq, update injection points to PreToolUse:Bash
- utils.mjs: drop dead knowledge/agents.system config + isSystemAgent
- delete stale docs flow.md/hooks.md/commands.md; rewrite file-tree.md + grepai.md
2026-06-29 18:53:40 +01:00
f41244072f
v3.16.1: full 3-level doc-sync + mobile overflow fix + alias-doc fix
...
Docs-only release. Reconciled skill/agent/hook counts to 33/16/9 across all three doc levels, fixed the provider-switch web-doc alias (claudedeepseek -> claudeds), fixed mobile horizontal-overflow on doc pages, and removed the final GLM/Z.ai residue.
2026-06-17 13:21:41 +01:00
7d80a8c8c4
text-optimize (deep+standard) + shorten skill/agent descriptions to caps; update generators
2026-06-11 19:22:15 +01:00
2aeff9a6d5
v3.4.16: merge install skill into setup Phase 0
...
Prerequisites check (brew, coreutils, jq, grepai) now runs automatically
as Phase 0 before project analysis. If all required tools are present,
Phase 0 is skipped silently.
2026-04-02 14:08:47 +01:00
5a6a0510c0
v3.4.14: brewtools plugin, teams skill, mode switcher, hook-creator v2.1.89+
...
brewtools (new plugin):
- text-optimize, text-human, secrets-scan skills + text-optimizer agent
- BT_PLUGIN_ROOT injection via session-start hook
brewcode:
- /brewcode:teams — dynamic agent team creation/management
- Mode Switcher system (modes/, getActiveMode(), hooks injection)
- plan: Dynamic Agent Resolution from .claude/teams/ roster
- skills: Step 2.5 mode-switcher detection
- skill-creator: Mode Switcher pattern docs
- hook-creator: v2.1.89+ (PermissionDenied event, defer, 26 events)
- removed: text-optimize, text-human, secrets-scan → brewtools
- convention P5: fallback when brewtools not installed
2026-04-02 08:22:13 +01:00
43953b56d7
v3.1.0
2026-02-28 20:01:40 +00:00
14b01756ee
v2.15.6
2026-02-21 11:53:11 +00:00
93f15bf5ef
v2.15.5
2026-02-20 07:45:52 +00:00
b33e0162fa
v2.15.4
2026-02-19 08:35:05 +00:00
kochetkov-ma