* feat(devops): add auto-evolution loop (PR review + BMAD bug/feature pipeline)
Extends the DevOps Hand to periodically scan configured GitHub repos and:
- review open PRs via the existing code-reviewer sub-agent, posting a
single COMMENT review back to GitHub (never auto-APPROVE)
- triage open issues via labels first, single-prompt LLM fallback
- dispatch actionable issues (bug-fix / feature) to a new implementer
sub-agent which runs the BMAD pipeline (Brainstorm -> Architect ->
PRD -> Implement) scaled by bmad_strictness and produces a DRAFT PR
Safety floor (always on):
- draft PRs only, never auto-ready, never merge
- never push to main/master/protected branches
- escalates to devops_queue.json when touching workspace Cargo.toml,
migrations, secrets, or >30 changed files
- 70% per-turn token budget cap so subsequent ticks have headroom
New settings: auto_evolve, evolution_repos, evolution_check_interval,
bmad_strictness. New sub-agent: agents.implementer. New SKILL.md
sections: Issue Triage Playbook, PR Review Automation, Bug Fix
Playbook, BMAD Feature Pipeline, Draft PR Creation. Three new
dashboard metrics: prs_reviewed, issues_processed, draft_prs_opened.
* fix(devops): address PR review — close blocking + medium + style issues
Blocking (5):
- add max_changed_files setting (was referenced in implementer prompt
but never defined)
- drop metering_query reference (tool isn't in tools = [...] list);
agent self-paces against budget instead
- fix \n\n literal in jq --arg for issue cross-link comment; compose
body in shell with printf so newlines survive
- resolve BASE_BRANCH via /repos/owner/repo .default_branch instead
of relying on an undefined variable
- complete reviewer-verdict → GitHub review-event mapping (4 cases,
not just request_changes); block routes through REQUEST_CHANGES
with a blocking-prefix in the body, approve downgrades to COMMENT
Medium (5):
- correct Phase 6 → Phase 7 in the auto-evolution settings comment
- remove schedule_create busy-loop confusion; Phase 7 fires per-turn
while the Hand is already frequency = "continuous", with cadence
enforced via devops_evolution_cursor memory key
- generalize the forbid-main-worktree wording — discover and honor
whatever pre-commit / pre-push / commit-msg hooks the upstream
repo configures (was librefang-specific)
- clarify the AI-attribution rule: ban LLM-vendor attribution
(Claude, GPT, 🤖, etc.) but allow process attribution
(DevOps Hand → implementer) for traceability
- add USER_TYPE = "Bot" short-circuit that was extracted but never
applied (bots get a token-cheap skip, not a deep review)
Style (2):
- document the four event_publish event names (devops_evolution_*)
in a new SKILL.md table alongside the memory-keys table
- justify implementer's max_history_messages = 100 with a comment
(BMAD 4 phases × cargo build/test chains needs headroom)
* docs(devops): tighten evolution snippets (D1-D4 second-review nits)
D1 -- show SUMMARY_BODY (and VERDICT) assignment in PR review snippet:
add explicit jq -r .summary / .verdict extraction from reviewer_output.json
so the agent reading SKILL.md doesn't have to infer where these come from.
D2 -- reword strict-mode wait semantics in both HAND.toml and SKILL.md:
'Stop. Wait...' was misleading because the agent loop has no in-turn
pause primitive. Now spells out: end the current turn after queueing,
let the continuous tick re-read the queue, resume on approved / skip
on pending / abandon on rejected. Explicitly forbids busy-wait and
sleep loops.
D3 -- restructure bot / huge-diff short-circuit so agent-tool calls are
expressed as numbered agent steps, not as '# memory_store ...' comments
inside a bash block. The bash block now only extracts cheap signals;
the decision and the tool calls are clearly agent-level.
D4 -- remove the misleading 'exit 0' from the short-circuit bash and
add a one-liner noting that exit 0 inside shell_exec only ends one
shell session, not the Phase 7 pass; the agent must choose to move on.
All 32 agent manifests and 17 hands shipped with empty mcp_servers /
skills lists, which the kernel interprets as "no filter" — every
globally-configured MCP server's tools and every installed skill get
injected into the prompt on every LLM call. On a typical instance (9
MCP servers, ~85 MCP tools + ~82 built-in tools) that's ~50k input
tokens per turn spent on definitions the agent never uses.
Changes
-------
32 agents/*/agent.toml:
- mcp_servers: 1-4 per agent. memory wherever state persists across
turns; fetch / exa-search / brave-search only where the prompt
actually calls for web; git / github / filesystem on engineering
agents; gmail / google-calendar / linear / jira on productivity
agents whose prompts mention them.
- skills: per-role allowlist driven by what the system_prompt names
(e.g. coder → rust/python/typescript/git/shell-scripting; devops-
lead → docker/kubernetes/terraform/ansible/ci-cd/helm/prometheus/
sysadmin). Generalists (assistant) keep skills = [] (see "Open
items" below).
- skills_disabled = true on the four short-conversational agents
(hello-world, recipe-assistant, health-tracker, home-automation).
Their system prompts never instruct the LLM to consult any skill,
so loading all 60 was pure waste. They also drop the explicit
max_history_messages override and inherit the kernel default (60).
- max_history_messages tiered by workload shape:
60 short conversational (hello-world, recipe, health-tracker,
home-automation) — inherits the rising kernel default
(`DEFAULT_MAX_HISTORY_MESSAGES = 60`); no override needed.
60 single-turn task agents (writer, translator, doc-writer,
email-assistant, customer-support, sales-assistant, recruit-
er, social-media, personal-finance, tutor, travel-planner,
meeting-assistant, ops, devops-lead, planner) — explicit
override at the same value to lock the cap if the kernel
default moves again.
80 multi-step / tool-heavy (coder, debugger, architect, code-
reviewer, test-engineer, security-auditor, analyst, data-
scientist, academic-researcher, researcher, legal-assistant)
120 coordinators (assistant, orchestrator) — long multi-agent
sessions where prompt-cache continuity is critical
All values sit at or above the kernel default. Pinning lower
would thrash the prompt cache (the failure mode #91 fixed for
the creator hand by *raising* the cap, not lowering it).
17 hands/*/HAND.toml:
- hand-level mcp_servers / skills now declared on every hand, so
every [agents.*] inside inherits a sensible allowlist.
- skills_disabled = true placed on each [agents.*] inside clip and
creator (pure media pipelines that don't benefit from any skill).
HandDefinitionRaw in librefang-hands does NOT have a top-level
skills_disabled field — declaring it at the hand top level would
be silently dropped by serde, so the setting must live on the
AgentManifest of each sub-agent role.
- devteam: expand existing mcp_servers = ["github"] to include
memory / git / filesystem; populate skills with the expected
dev-team expertise (replacing the placeholder skills = []).
- wiki: replace placeholder mcp_servers = [] with [memory, fetch,
filesystem]. Hand-level skills stays [].
- lead: hand-level skills was originally [email-writer, writing-
coach, interview-prep]; interview-prep is for job-interview
preparation, not lead generation. Replaced with data-analyst
(used by the qualification-scoring step in the prompt).
schema.toml: register mcp_servers / skills / max_history_messages on
the agent field schema so machine consumers (RegistrySchema in
librefang-types) see the new top-level fields. The
max_history_messages description now points at
librefang_runtime::agent_loop::DEFAULT_MAX_HISTORY_MESSAGES (60
today) by name, so the schema doesn't go stale when the constant
moves again.
agents/README.md: example block + "Adding a New Agent" checklist
mention the allowlists; max_history_messages example is shown
commented out with a prompt-cache caveat.
Open items
----------
`assistant` (the default user-facing agent) keeps `skills = []`
deliberately. It is the generalist entry point — capping its skill
surface at a small allowlist would defeat its "delegate to any
specialist" job. The trade-off is that this single agent still pays
the full skill-definition load on every turn; operators who want a
strict allowlist for `assistant` can override it after install.
Why not adopt PR #89's approach
-------------------------------
#89 covers similar ground but with three issues this PR avoids:
1. mcp_servers = ["_none"] sentinel. #89's body explicitly notes
it's pending upstream librefang#4808 (mcp_disabled). Shipping a
magic-string today means coming back later to clean it up. This
PR uses real allowlists.
2. max_history_messages = 8 / 12 / 15 / 20. Far below today's
kernel default (60) and #91's direction for long-workflow hands
(80–120). Every turn that hits the cap invalidates the cached
prompt prefix; the cost of cache misses exceeds the saving from
shorter history. This PR uses 60–120.
3. Doubling max_llm_tokens_per_hour (coder 200k→500k, assistant
300k→500k) widens the per-agent budget — the opposite direction
from #87's "reduce per-call cost" goal. Left to the operator's
instance-specific tuning.
Refs librefang/librefang-registry#87, librefang/librefang-registry#89
* fix(creator): raise max_history_messages to 80 for polling workflows
Creator Hand's async video_generate path polls video_status every 15-20s
until completion (1-3 min typical), consuming ~5-15 turns per video
request. Combined workflows (video + TTS + music) plus normal back-and-
forth cross the kernel default of 40 messages quickly, which surfaced
in user logs as:
WARN run_agent_loop: Trimming old messages at safe turn boundary
agent=creator:creator-hand total_messages=41 trimming=2
INFO run_agent_loop: prompt cache metrics for turn
hit_ratio=0.0 creation=0 read=0
Every turn was hitting the trim cap and invalidating the prompt-cache
prefix. 80 covers ~30 polling iterations plus a comfortable pre-context
window without runaway memory growth. Other hands keep the default 40.
* ci(refresh-cache): open PR instead of pushing directly to main
Branch protection on `main` started rejecting the workflow's auto-commit
with GH006 "Changes must be made through a pull request" — see run
25632824585 on 2026-05-10 against commit 6785807 (the first push that
hit the tightened protection). Direct push is precisely what the file's
own security comment (#1) warns against ("Compromised maintainer pushes
a malicious plugins-index.json directly to main. Mitigation: GitHub
branch protection on main requires PR review"), so the fix preserves
that gate rather than working around it.
The workflow now creates a short-lived `automation/refresh-indexes-<sha>`
branch, commits the regen there, pushes, and opens a PR back to main
via `gh pr create`. Maintainers see a one-click squash-merge.
Permissions: add `pull-requests: write` to the existing `contents: write`
so `gh pr create` can be authorised through the default GITHUB_TOKEN.
The post-merge run on the index PR is a no-op (no diff under
`hands/**`, `plugins/**`, etc. between consecutive states), so no
`[skip ci]` marker is needed and no loop is possible.
Without this fix, every content PR landing on main leaves
plugins-index.json + registry-index.json stale, blocking new agents and
hands from reaching daemons until a maintainer manually regenerates.
* fix(hands): raise max_history_messages on long-workflow coordinators
Three hand coordinators have workflows that routinely exceed the kernel
default history cap on a single user turn:
- researcher (max_iterations=80) — deep web_search → web_fetch →
summarize loops with multi-source synthesis. 80 iterations × ~4
messages each → 200+ messages per user turn. Set to 120.
- devops (max_iterations=60) — incident response and CI/CD fan out
into long shell_exec chains (logs, retries, post-mortems). Set to 80.
- predictor (max_iterations=60) — long reasoning chains accumulating
signals across many web/knowledge queries, with scheduled re-checks
referring back. Set to 80.
Creator's existing override is rephrased "raise above the kernel
default" so the comment stays correct regardless of the order this PR
and the upstream kernel-default bump (librefang side) land in.
Other hands (lead/linkedin/reddit/clip/analytics/apitester/browser/
collector/strategist) stay on the kernel default; the upstream bump
covers them.
* refactor: migrate icon fields from emoji to lucide:<name> tokens
Every TOML manifest's `icon = "<emoji>"` line is replaced with
`icon = "lucide:<kebab-name>"` — a reference to a lucide-react icon,
which the librefang.ai site and dashboard render as crisp SVG. Reasons
for the switch:
- Emoji render very differently across OS/browser/font stacks; the
registry catalog looked inconsistent from one row to the next.
- Five manifests (clip / creator / linkedin / reddit / twitter) had
their icons stored as literal Python-style escape strings
("\\U0001F3AC") because the TOML parser upstream never decoded
them. Switching away from emoji drops that class of bug entirely.
- As a drive-by, also decode the \\uXXXX accent escapes in the
[i18n.fr] block of hands/creator/HAND.toml so "Créateur" shows
up correctly.
87 files touched. example manifests left untouched (still "TODO").
* fix: backfill i18n name + drop the single-member email category
- Every existing [i18n.<lang>] block now has a `name` field. 60 files
previously translated description but kept the English name
implicitly — which rendered as "some English some Chinese" in the
registry UI. Fill in the missing name from the English brand (or a
known localized equivalent: DingTalk→钉钉, Feishu→飞书, Email→
电子邮件 / メール / E-Mail / Correo / Courriel, and a handful of
hands that have Chinese product names like 视频剪辑 Hand).
- channels/email.toml was the only item under category="email";
reclassify it as "messaging" so the sub-category filter chip list
on the category page isn't littered with singletons.
* feat(i18n): localize 76 agents/integrations/plugins into 7 languages
Adds full [i18n.zh], [i18n.zh-TW], [i18n.ja], [i18n.ko], [i18n.de],
[i18n.es], [i18n.fr] blocks with name + description to every manifest
that previously shipped English-only.
Coverage:
- 32 agents (academic-researcher, analyst, architect, assistant,
code-reviewer, coder, customer-support, data-scientist, debugger,
devops-lead, doc-writer, email-assistant, health-tracker,
hello-world, home-automation, legal-assistant, meeting-assistant,
ops, orchestrator, personal-finance, planner, recipe-assistant,
recruiter, researcher, sales-assistant, security-auditor,
social-media, test-engineer, translator, travel-planner, tutor,
writer)
- 33 integrations (AWS, Azure, Bitbucket, Brave Search, Discord,
Dropbox, Elasticsearch, Exa Search, Fetch, Filesystem, GCP, Git,
GitHub, GitLab, Gmail, Google Calendar, Google Drive, Google Maps,
Jira, Linear, Memory, MongoDB, Notion, PostgreSQL, Puppeteer, Redis,
Sentry, Sequential Thinking, Slack, SQLite, Teams, Time, Todoist) —
brand names kept as-is across all locales, only descriptions
translated.
- 11 plugins (auto-summarizer, context-decay, conversation-logger,
episodic-memory, guardrails, keyword-memory, mempalace-indexer,
sentiment-tracker, todo-tracker, topic-memory, user-profile)
The descriptions are one-line summaries — hand-translated rather than
machine-generated, so technical terms (MCP, PR, CI/CD, etc.) stay
consistent across locales.
* feat(i18n): close remaining per-lang gaps for channels, workflows, devteam
Third pass on i18n coverage. Every non-example manifest now carries a
full set of [i18n.zh], [i18n.zh-TW], [i18n.ja], [i18n.ko], [i18n.de],
[i18n.es], [i18n.fr] blocks.
- 44 channel adapters: added French descriptions (zh/zh-TW/ja/ko/de/es
were already present). Brand names kept as-is in all locales so users
recognize Discord / Slack / LINE / etc. consistently.
- 22 workflows: filled zh-TW / ja / ko / de / es / fr blocks. Each
translation mirrors the existing zh one in structure and tone so the
catalog reads consistently across locales.
- hands/devteam/HAND.toml: added the four langs that were missing
(zh-TW, de, es, fr).
Only the 6 templates under examples/ are left without i18n blocks on
purpose — they still contain "TODO:" placeholders.
Add [i18n.zh.agents.*] sections to all 15 HAND.toml files,
providing Chinese translations for agent names and descriptions.
Total: 51 agent translations across 15 hands.
Add [i18n.zh], [i18n.zh-TW], [i18n.ja], [i18n.ko], [i18n.de], [i18n.es]
sections with translated descriptions to:
- 15 Hand TOML files (hands/*/HAND.toml)
- 44 Channel TOML files (channels/*.toml)
This enables the website to display localized Hand and Channel descriptions
based on the user's selected language.
- Fix French accent characters (é/è/ê/ç/â/ô) across all 14 HAND.toml files
- Fix German special characters (ä/ö/ü/ß) across all 14 HAND.toml files
- Add category translations to all 6 i18n language blocks in all 14 hands
- Enhance SKILL.md content for 9 hands with practical examples and workflows
- Trim bloated SKILL.md files (apitester 1400→892, devops 1301→870)
- Rewrite root README.md with accurate stats, complete hand/integration tables
- Update hands/README.md with full 14-hand listing and i18n documentation
- Add version = "1.0.0" to all 14 hands
- Add approval_mode toggle to clip and devops hands (queue actions
for user review before executing)
- Strengthen routing aliases for analytics, apitester, browser,
collector, devops, lead, predictor, researcher, strategist
* chore: remove router agent
builtin:router has been replaced by LLM intent routing in the kernel.
Assistant is now the sole entry point — see librefang/librefang#1336.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* style: format all TOML files with taplo
Fix CI taplo format check by running `taplo fmt` on all 132 TOML files.
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Evan