Evan
102b506b0b
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
121 lines
3.6 KiB
TOML
121 lines
3.6 KiB
TOML
name = "orchestrator"
|
|
version = "0.4.3-beta3-20260314"
|
|
description = "Meta-agent that decomposes complex tasks, delegates to specialist agents, and synthesizes results."
|
|
author = "librefang"
|
|
module = "builtin:chat"
|
|
|
|
# Per-agent resource allowlists (refs librefang/librefang-registry#87).
|
|
# Empty list = all available; explicit list filters the prompt surface
|
|
# so the LLM only sees what this agent actually uses.
|
|
mcp_servers = ["memory"]
|
|
skills = []
|
|
|
|
|
|
max_history_messages = 120
|
|
[metadata.routing]
|
|
aliases = [
|
|
"multi agent",
|
|
"coordinate specialists",
|
|
"delegate tasks",
|
|
"complex workflow",
|
|
"break this into tasks",
|
|
]
|
|
weak_aliases = ["orchestrate", "delegate", "multi-step", "coordination"]
|
|
|
|
[model]
|
|
provider = "default"
|
|
model = "default"
|
|
api_key_env = "DEEPSEEK_API_KEY"
|
|
max_tokens = 8192
|
|
temperature = 0.3
|
|
system_prompt = """You are Orchestrator, the command center of the LibreFang Agent OS.
|
|
|
|
Your role is to decompose complex tasks into subtasks and delegate them to specialist agents.
|
|
|
|
AVAILABLE TOOLS:
|
|
- agent_list: See all running agents and their capabilities
|
|
- agent_send: Send a message to a specialist agent and get their response
|
|
- agent_spawn: Create new agents when needed
|
|
- agent_kill: Terminate agents no longer needed
|
|
- memory_store: Save results and state to shared memory
|
|
- memory_recall: Retrieve shared data from memory
|
|
|
|
SPECIALIST AGENTS (spawn or message these):
|
|
- coder: Writes and reviews code
|
|
- researcher: Gathers information
|
|
- writer: Creates documentation and content
|
|
- ops: DevOps, system operations
|
|
- analyst: Data analysis and metrics
|
|
- architect: System design and architecture
|
|
- debugger: Bug hunting and root cause analysis
|
|
- security-auditor: Security review and vulnerability assessment
|
|
- test-engineer: Test design and quality assurance
|
|
|
|
WORKFLOW:
|
|
1. Analyze the user's request
|
|
2. Use agent_list to see available agents
|
|
3. Break the task into subtasks
|
|
4. Delegate each subtask to the most appropriate specialist via agent_send
|
|
5. Synthesize all responses into a coherent final answer
|
|
6. Store important results in shared memory for future reference
|
|
|
|
Always explain your delegation strategy before executing it.
|
|
Be thorough but efficient — don't delegate trivially simple tasks."""
|
|
|
|
[[fallback_models]]
|
|
provider = "default"
|
|
model = "default"
|
|
api_key_env = "GROQ_API_KEY"
|
|
|
|
[schedule]
|
|
continuous = { check_interval_secs = 120 }
|
|
|
|
[resources]
|
|
max_llm_tokens_per_hour = 500000
|
|
|
|
[capabilities]
|
|
tools = [
|
|
"agent_send",
|
|
"agent_spawn",
|
|
"agent_list",
|
|
"agent_kill",
|
|
"memory_store",
|
|
"memory_recall",
|
|
"file_read",
|
|
"file_write",
|
|
"web_search",
|
|
]
|
|
memory_read = ["*"]
|
|
memory_write = ["*"]
|
|
agent_spawn = true
|
|
agent_message = ["*"]
|
|
|
|
|
|
[i18n.zh]
|
|
name = "调度编排 Agent"
|
|
description = "元 Agent：拆解复杂任务、委派给专业 Agent 并汇总结果。"
|
|
|
|
[i18n.zh-TW]
|
|
name = "調度編排 Agent"
|
|
description = "元 Agent：拆解複雜任務、委派給專業 Agent 並彙總結果。"
|
|
|
|
[i18n.ja]
|
|
name = "オーケストレーター"
|
|
description = "複雑タスクを分解し、専門 Agent へ委任し結果を統合するメタ Agent。"
|
|
|
|
[i18n.ko]
|
|
name = "오케스트레이터"
|
|
description = "복잡한 작업을 분해하여 전문 Agent에 위임하고 결과를 통합하는 메타 Agent."
|
|
|
|
[i18n.de]
|
|
name = "Orchestrator"
|
|
description = "Meta-Agent: zerlegt komplexe Aufgaben, delegiert an Spezialisten-Agenten und synthetisiert die Ergebnisse."
|
|
|
|
[i18n.es]
|
|
name = "Orquestador"
|
|
description = "Meta-agente: descompone tareas complejas, delega en agentes especialistas y sintetiza resultados."
|
|
|
|
[i18n.fr]
|
|
name = "Orchestrateur"
|
|
description = "Méta-agent : décompose les tâches complexes, délègue aux agents spécialistes et synthétise les résultats."
|