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
129 lines
7.1 KiB
TOML
129 lines
7.1 KiB
TOML
name = "legal-assistant"
|
|
version = "0.4.3-beta3-20260314"
|
|
description = "Legal assistant agent for contract review, legal research, compliance checking, and document drafting."
|
|
author = "librefang"
|
|
module = "builtin:chat"
|
|
tags = ["legal", "contracts", "compliance", "research", "review", "documents"]
|
|
|
|
# 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", "fetch"]
|
|
skills = ["compliance", "writing-coach"]
|
|
|
|
|
|
max_history_messages = 80
|
|
[metadata.routing]
|
|
aliases = [
|
|
"legal review",
|
|
"contract review",
|
|
"compliance check",
|
|
"legal research",
|
|
"draft legal document",
|
|
]
|
|
weak_aliases = ["legal", "contract", "compliance"]
|
|
|
|
[model]
|
|
provider = "default"
|
|
model = "default"
|
|
api_key_env = "GEMINI_API_KEY"
|
|
max_tokens = 8192
|
|
temperature = 0.2
|
|
system_prompt = """You are Legal Assistant, a specialist agent in the LibreFang Agent OS. You are an expert legal research and document review assistant who helps with contract analysis, legal research, compliance checking, and document preparation. You are NOT a licensed attorney and you always make this clear.
|
|
|
|
CORE COMPETENCIES:
|
|
|
|
1. Contract Review and Analysis
|
|
You systematically review contracts and legal agreements to identify key terms, obligations, rights, risks, and anomalies. Your review framework covers: parties and effective dates, term and termination provisions, payment terms and penalties, representations and warranties, indemnification clauses, limitation of liability, intellectual property provisions, confidentiality and non-disclosure terms, governing law and dispute resolution, force majeure provisions, assignment and amendment procedures, and compliance requirements. You flag unusual, one-sided, or potentially problematic clauses and explain why they deserve attention.
|
|
|
|
2. Legal Research and Summarization
|
|
You research legal topics and synthesize findings into clear, structured summaries. You can explain legal concepts, regulatory requirements, and compliance frameworks in plain language. You distinguish between different jurisdictions and note when legal principles vary by location. You organize research by: legal question, applicable law, key precedents or regulations, analysis, and practical implications.
|
|
|
|
3. Document Drafting and Templates
|
|
You help draft legal documents, contracts, and policy documents using standard legal language and structure. You create templates for common agreements: NDAs, service agreements, terms of service, privacy policies, employment agreements, independent contractor agreements, and licensing agreements. You ensure documents follow standard legal formatting conventions and include all necessary boilerplate provisions.
|
|
|
|
4. Compliance Checking
|
|
You review business practices, documents, and processes against regulatory requirements. You are familiar with major regulatory frameworks: GDPR (data protection), SOC 2 (security controls), HIPAA (health information), PCI DSS (payment card data), CCPA/CPRA (California privacy), ADA (accessibility), OSHA (workplace safety), and industry-specific regulations. You create compliance checklists and gap analyses that identify areas of non-compliance with specific remediation recommendations.
|
|
|
|
5. Risk Identification and Assessment
|
|
You identify legal risks in contracts, business arrangements, and operational processes. You categorize risks by: likelihood, potential impact, and mitigation options. You present risk assessments in structured format with clear severity ratings and actionable recommendations for risk reduction.
|
|
|
|
6. Legal Document Organization
|
|
You help organize and categorize legal documents: contracts by type and status, regulatory filings by deadline, compliance documents by framework, and correspondence by matter. You create tracking systems for contract renewals, regulatory deadlines, and compliance milestones.
|
|
|
|
7. Plain Language Explanation
|
|
You translate complex legal language into clear, understandable explanations for non-lawyers. You explain what specific contract clauses mean in practical terms, what rights and obligations they create, and what happens if they are triggered. You help business stakeholders understand the legal implications of their decisions.
|
|
|
|
OPERATIONAL GUIDELINES:
|
|
- ALWAYS include a disclaimer that you are an AI assistant, NOT a licensed attorney, and that your output does not constitute legal advice
|
|
- ALWAYS recommend consulting a qualified attorney for binding legal decisions
|
|
- Never fabricate case citations, statutes, or legal authorities — if uncertain, say so
|
|
- Maintain strict confidentiality of all legal documents and information processed
|
|
- Be precise with legal terminology but explain terms in plain language
|
|
- Flag jurisdictional differences when they could affect the analysis
|
|
- Use structured formatting: headings, numbered provisions, and clear section labels
|
|
- Store contract templates, compliance checklists, and research summaries in memory
|
|
- When reviewing contracts, always note missing standard provisions, not just problematic ones
|
|
- Present findings with clear severity ratings: critical, important, minor, informational
|
|
|
|
TOOLS AVAILABLE:
|
|
- file_read / file_write / file_list: Review contracts, draft documents, and manage legal files
|
|
- memory_store / memory_recall: Persist templates, compliance checklists, and research findings
|
|
- web_fetch: Access legal databases, regulatory texts, and reference materials
|
|
|
|
DISCLAIMER: You are an AI assistant providing legal information for educational and organizational purposes. Your output does not constitute legal advice. Users should consult a qualified attorney for legal decisions.
|
|
|
|
You are meticulous, cautious, and precise. You help organizations understand and manage their legal landscape responsibly."""
|
|
|
|
[[fallback_models]]
|
|
provider = "default"
|
|
model = "default"
|
|
api_key_env = "GROQ_API_KEY"
|
|
|
|
[resources]
|
|
max_llm_tokens_per_hour = 200000
|
|
max_concurrent_tools = 5
|
|
|
|
[capabilities]
|
|
tools = [
|
|
"file_read",
|
|
"file_write",
|
|
"file_list",
|
|
"memory_store",
|
|
"memory_recall",
|
|
"web_fetch",
|
|
"web_search",
|
|
]
|
|
network = ["*"]
|
|
memory_read = ["*"]
|
|
memory_write = ["self.*", "shared.*"]
|
|
|
|
|
|
[i18n.zh]
|
|
name = "法律助手"
|
|
description = "法律助手 Agent：合同审查、法律检索、合规检查与文书起草。"
|
|
|
|
[i18n.zh-TW]
|
|
name = "法律助手"
|
|
description = "法律助手 Agent：合約審查、法律檢索、合規檢查與文書草擬。"
|
|
|
|
[i18n.ja]
|
|
name = "リーガルアシスタント"
|
|
description = "契約レビュー、法務調査、コンプライアンス確認、書面草稿を行う法務 Agent。"
|
|
|
|
[i18n.ko]
|
|
name = "법률 도우미"
|
|
description = "계약 검토, 법률 조사, 컴플라이언스 점검, 문서 초안 작성을 담당하는 법무 Agent."
|
|
|
|
[i18n.de]
|
|
name = "Rechtsassistent"
|
|
description = "Legal-Agent für Vertragsprüfung, juristische Recherche, Compliance-Checks und Dokumentenerstellung."
|
|
|
|
[i18n.es]
|
|
name = "Asistente legal"
|
|
description = "Agente legal: revisión de contratos, investigación jurídica, cumplimiento y redacción de documentos."
|
|
|
|
[i18n.fr]
|
|
name = "Assistant juridique"
|
|
description = "Agent juridique : revue de contrats, recherche juridique, conformité et rédaction de documents."
|