Pair with the worker-side simplification on the librefang PR — the
worker is now a pure transport (no key material, no signing) and this
repo's CI takes over signature production.
scripts/sign-plugins-index.mjs reads REGISTRY_PRIVATE_KEY from a GitHub
Actions secret, signs plugins-index.json with Ed25519, and writes
plugins-index.json.sig alongside it. Aborts loudly when the secret is
missing so a misconfigured CI can't silently ship an unsigned payload.
The workflow now runs build → sign → commit (.json + .sig) → push →
poke worker /refresh. The worker fetches the committed .json + .sig
verbatim and stores both — the daemon then verifies against the
embedded pubkey it ships with.
Closes PR review CRITICAL #1: the worker is no longer a sign-anything
oracle reachable via REGISTRY_REFRESH_TOKEN. Trust root is now this
repo's branch protection + Actions secret scope, not a token any CI
job that can talk to stats.librefang.ai can use to mint signatures.
Note: the keypair was rotated as part of this change (PR not yet
merged so no daemon TOFU pins exist). New pubkey:
ClGa0Ucap8NdrKAy1rw9Tt6A9I8eg4zJ53+xIuKMuq0=
The plugins-index.json.sig committed here is signed with the matching
new private key, in lockstep with the daemon EMBEDDED_REGISTRY_PUBKEY
constant and all three worker [vars] entries.
Pair with the existing plugins-index.json path (which feeds the
daemon's signed install lane). registry-index.json mirrors the
dict-shaped payload the registry-worker's cron currently builds via
40+ GitHub Contents API calls — but built locally from the checked-out
tree by scripts/build-registry-index.mjs, so the worker only fetches
ONE file per category type (2 total: plugins + registry) on refresh.
Workflow now picks up content changes across all 8 category dirs
(was: plugins/ only) so dashboard updates land within seconds of a
push instead of waiting for the 02:00 UTC cron tick.
The dashboard's /api/registry endpoint reads kv_store('registry_data');
the worker's forced-refresh now writes that key with these bytes and
purges the Cache-API entry, so the next dashboard hit sees fresh
data instead of the 1h-cached previous payload.
Generated counts on first build: 11p 17h 32a 60s 44c 57pr 22w 33mcp.
Walking ~40+ plugin TOMLs via the GitHub Contents API from the worker
exceeded the Workers Free 50-subrequest-per-invocation limit, leaving
the daemon's signed plugins index either empty or partial after every
forced refresh.
Move the walk into the repo: scripts/build-plugins-index.mjs reads each
plugins/<name>/plugin.toml directly from the checked-out tree and emits
a sorted flat array (name, version?, description?, needs?) at
plugins-index.json. The CI workflow regenerates and commits this file
on every push under plugins/, then pokes the worker's
/api/registry/refresh — which now fetches the single committed
plugins-index.json (1 subrequest), validates the JSON shape, and
re-signs it with Ed25519. Refresh cost is now constant in registry
size, not linear.
The dashboard's dict-shaped /api/registry payload is unchanged — that
still rebuilds via the daily 02:00 UTC cron.
Extend modality enum to support video and music, then register the
non-text MiniMax models that were already declared in
media_capabilities but had no concrete entries:
- image-01 ($0.0035/image)
- speech-2.8/2.6 hd & turbo ($60-$100 per 1M chars)
- Hailuo 2.3 Fast / 2.3 / 02 video models ($0.10-$0.56 per video)
- music-2.6, lyrics_generation
Per-call pricing is documented in inline comments since the schema's
token-based cost fields don't naturally fit per-call billing.
schema.toml and scripts/validate.py both updated; the change is
additive (existing modality values remain valid).
Introduces image-generation models as a first-class [[models]] entry via
a new `modality` field on the model schema ("text" default, "image",
"audio"). When modality != "text", context_window / max_output_tokens
are optional since no conventional context gate exists — OpenAI's
gpt-image-2 docs omit them.
Adds `image_input_cost_per_m` / `image_output_cost_per_m` alongside
existing text token cost fields to cover the 4-price structure OpenAI
uses for image generation (text $5/$10, image $8/$30 per 1M tokens).
Validator updated to:
- accept any modality in {text, image, audio}
- require context_window/max_output_tokens only for modality=text
- range-check the two new cost fields
gpt-image-2 entry added to providers/openai.toml with pricing sourced
from https://developers.openai.com/api/docs/pricing. Snapshot
gpt-image-2-2026-04-21 listed as alias.
* fix(providers): remove ~anthropic, skip ~ prefixes in sync script
OpenRouter uses ~ prefixes for internal auto-routing aliases (e.g. ~anthropic).
These are not real providers — they already route through openrouter.toml.
The generated ~anthropic.toml was confusing (looked like a stale backup)
and redundant with the existing openrouter provider.
- Delete providers/~anthropic.toml
- Skip provider IDs starting with ~ in sync-pricing.py --create-missing
* fix(providers): remove morph, aider, kwaipilot
- morph: specialized code-editing/patching tool, not a general LLM provider
- aider: CLI meta-tool wrapper (base_url empty), redundant with claude-code/codex-cli/gemini-cli/qwen-code
- kwaipilot: Kwai internal coding assistant routed via OpenRouter, niche
* fix(sync): add morph/aider/kwaipilot to SKIP_PROVIDERS to prevent re-creation
* feat(sync): merge OpenRouter-only providers into openrouter.toml
Instead of generating standalone .toml files that just wrap the OpenRouter
endpoint, merge their models directly into openrouter.toml with the
standard 'openrouter/{provider}/{model}' ID convention.
- Add _build_model_fields() and _model_lines() helpers to deduplicate
model rendering between standalone and merged paths
- Add merge_into_openrouter() that appends new models idempotently
- generate_provider_toml() now only runs for providers in PROVIDER_API
- --create-missing routes OpenRouter-only providers to merge_into_openrouter
* fix(providers): remove 14 OpenRouter-only standalone files
These providers have no direct public API and all route through
openrouter.ai/api/v1. Per the new sync-pricing.py policy, their models
will be merged into openrouter.toml on the next CI run instead of
living in separate files that just wrap the OpenRouter endpoint.
Removed: allenai, deepcogito, essentialai, inclusionai, inflection,
liquid, meituan, nex-agi, nousresearch, prime-intellect, relace,
switchpoint, tngtech, writer
* fix(providers): remove 7 niche providers with no driver support
No dedicated LLM driver code exists for these providers — they rely
purely on OpenAI-compatible passthrough with no special handling.
Removing them reduces registry noise; users can still reach them via
openrouter.toml if needed.
Removed: microsoft, ibm-granite, xiaomi, upstage, inception, aion-labs, arcee-ai
* fix(providers): remove ai21, chutes, venice
All three use ApiFormat::OpenAI with no special handling — pure passthrough.
No registry entry needed; users can reach them via openrouter.toml or by
adding a custom provider.
* docs(providers): rewrite README with full provider catalog and inclusion criteria
- List all 46 providers grouped by category with descriptions
- Document why each provider exists (direct API, unique endpoint, dedicated driver, local, CLI)
- Add inclusion criteria section explaining when to create standalone files vs merging into openrouter.toml
- Document sync script routing logic
- Update model counts: 49→46 providers, 339→232 models
* docs: add comprehensive READMEs for all registry sections + deepinfra provider
- agents/README.md: 32 agents across 7 categories with capability field reference
- channels/README.md: 44 channels across 5 categories with protocol reference table
- hands/README.md: 18 hands across 5 categories with HAND.toml format guide
- mcp/README.md: 33 MCP servers across 5 categories with transport/auth format
- plugins/README.md: 12 plugins with hook protocol documentation
- skills/README.md: 60 skills across 9 categories with SKILL.md format guide
- providers/deepinfra.toml: add DeepInfra serverless inference (5 models)
Standardize on Claude Code's SKILL.md format as every skill's source of
truth. skill.toml becomes an optional metadata layer for runtime, input
schema, and versioning — never for the prompt body.
- validate.py: require SKILL.md in every skill dir; when skill.toml also
exists, cross-check name/description consistency to prevent drift
- Add SKILL.md to the two custom-skill examples
- Move the meeting-agenda prompt body out of skill.toml into SKILL.md
- Rewrite skills/README.md to document the md-first, toml-as-metadata convention
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Claude Code-style skills use SKILL.md with YAML frontmatter instead of
skill.toml. Validator now accepts either form, unblocking the 60 bundled
skills restored in #42.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Providers with known public APIs use their official endpoints:
- meta-llama → api.llama.com/v1
- microsoft → models.inference.ai.azure.com (GitHub Models)
- ibm-granite → us-south.ml.cloud.ibm.com/ml/v1 (watsonx)
- tencent → api.hunyuan.cloud.tencent.com/v1
- morph → api.morphllm.com/v1
16 remaining providers without known public APIs route through
OpenRouter (base_url = openrouter.ai/api/v1, OPENROUTER_API_KEY).
sync-pricing.py updated with PROVIDER_API mapping.
Providers without their own public API now use OpenRouter as their
base_url with OPENROUTER_API_KEY, making them testable and usable
when the user has an OpenRouter key configured.
- 20 OpenRouter-only providers: set base_url to openrouter.ai/api/v1
- morph: set correct official API (api.morphllm.com/v1)
- sync-pricing.py: default to OpenRouter routing for new providers
- Merge unique models from duplicate providers into their hand-written
counterparts and remove the duplicates:
- alibaba (tongyi-deepresearch) → qwen
- amazon (nova-2-lite, nova-micro, nova-premier) → bedrock
- bytedance (ui-tars) → volcengine
- nvidia (nemotron-3-nano, nemotron-3-super, etc.) → nvidia-nim
- rekaai (reka-flash-3) → reka
- Set correct official API base_url for providers with public APIs:
arcee-ai, inception, morph, reka, upstage
- Set key_required=false for 20 providers only accessible through
hosting platforms (no public API)
- Update sync-pricing.py with SKIP_DUPLICATES, PROVIDER_API mapping,
and default key_required=false for future auto-generated providers
- Replace tier "free" with "fast" (valid tiers: frontier/smart/balanced/fast/local)
- Remove version suffix from teams-mcp integration id field
- Update sync-pricing.py to not generate invalid tier values
* fix: pin npm package versions in MCP integration templates
Prevent supply chain attacks by pinning exact versions instead of
using unpinned `npx -y @package` which pulls latest on every run.
23 of 25 integrations pinned. sqlite-mcp and aws skipped (packages
not found on npm registry).
* fix: use stable azure/mcp version instead of beta
* feat: add pricing sync script and update model prices from OpenRouter API
- scripts/sync-pricing.py fetches real-time pricing from OpenRouter
- Updated 64 price fields across 13 provider files
- Run periodically or in CI to keep prices current
* fix(validate): check for [agents] instead of [agent] in HAND.toml
All 14 hands use [agents.main] (plural) for multi-agent config,
but the validator was checking for [agent] (singular), causing
all hands to fail validation.
* fix(routing): resolve 19 routing alias collisions
Agent is a sub-unit of hand, so hands take priority for routing.
Remove conflicting aliases from agent side when hand already owns them.
- analyst: remove data analysis, analyze data, dashboard (owned by hand/analytics)
- data-scientist: remove statistical analysis, forecast, prediction (owned by hand/analytics, hand/predictor)
- sales-assistant: remove prospecting, sales, pipeline (owned by hand/lead, hand/devops)
- devops-lead: remove incident response, kubernetes, terraform (owned by hand/devops)
- researcher: remove deep research, research, literature review (owned by hand/researcher)
- academic-researcher: remove literature review, systematic review (owned by hand/researcher)
- social-media: remove duplicate content calendar from weak_aliases
- hand/collector: remove competitive analysis (owned by hand/strategist)
Each directory (agents, hands, integrations, plugins, providers,
scripts, skills) now has a README documenting its TOML format,
current contents, and contribution steps.
Community-maintained TOML catalog for LibreFang. New models can be added
via PR without requiring a LibreFang binary release.
Includes validation script, bilingual docs, and GitHub templates.
Evan Hu