Context-efficient UK pensions legal intelligence via CLI. Use for semantic search, gap analysis, citation chains, and handbook guidance. Triggers on: obligation search, compliance gap, citation, handbook guidance, trustee duties.
From 45blacknpx claudepluginhub 45black/uk-legal-pluginsThis skill uses the workspace's default tool permissions.
Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Details PluginEval's skill quality evaluation: 3 layers (static, LLM judge), 10 dimensions, rubrics, formulas, anti-patterns, badges. Use to interpret scores, improve triggering, calibrate thresholds.
Context-efficient access to UK pensions regulatory data via the apex-intel CLI. Use this skill when you need quick, targeted queries without loading multiple MCP servers.
Trigger conditions:
Prefer this over pensions-research skill when:
Keywords:
~/Projects/pensions/apex-governance/tools/apex_intel.py
Run with: uv run apex_intel.py <command> [options]
Find obligations using natural language:
uv run apex_intel.py search "trustee disclosure duties"
uv run apex_intel.py search "ESG investment governance" --scheme-type DB
uv run apex_intel.py search "funding requirements" --limit 5 --json
Options:
--limit, -l (int): Maximum results (1-50, default: 10)--scheme-type, -s: Filter by scheme (DB, DC, Hybrid, PSPS, All)--type, -t: Filter by category (Governance, Reporting, etc.)--json: Output as JSON for agent consumptionGet complete obligation with enrichment:
uv run apex_intel.py detail TPR-GC-001
uv run apex_intel.py detail PA04-S227-001 --json
Returns:
Identify missing obligations:
uv run apex_intel.py gaps --known TPR-GC-001,TPR-GC-002
uv run apex_intel.py gaps --known TPR-GC-001 --threshold 60 --priority High
Options:
--known, -k (required): Comma-separated obligation IDs--threshold, -t: Similarity threshold (0-100, default: 50)
--limit, -l: Max gaps to return--priority, -p: Filter by priority (High, Medium, Low, Critical)Find semantically similar obligations:
uv run apex_intel.py similar PA04-S227-001
uv run apex_intel.py similar TPR-GC-001 --limit 10 --json
Search handbook guidance:
uv run apex_intel.py handbook "integrated risk management"
uv run apex_intel.py handbook "ESG due diligence" --limit 10
Generate traceable citations:
uv run apex_intel.py cite TPR-GC-001
uv run apex_intel.py cite PA04-S227-001 --format full
uv run apex_intel.py cite TPR-GC-001 --format legal
Formats:
short: ID and source act onlyfull: Complete with implementing SIslegal: Formal legal citation# 1. Search for relevant obligations
uv run apex_intel.py search "disclosure duties" --limit 5 --json
# 2. Get details on most relevant
uv run apex_intel.py detail <obligation_id> --json
# 1. List obligations user knows about
# (from their input or previous queries)
# 2. Run gap analysis
uv run apex_intel.py gaps --known OB-001,OB-002,OB-003 --threshold 50 --json
# 3. Report gaps with citations
# 1. Get obligation details
uv run apex_intel.py detail TPR-GC-001 --json
# 2. Build citation chain from response
# Obligation → Source Act s.Section → Implementing SI → Enforcement Body
# 1. Semantic search
uv run apex_intel.py search "topic" --json
# 2. Get similar obligations for top result
uv run apex_intel.py similar <top_id> --json
# 3. Search handbook for guidance
uv run apex_intel.py handbook "topic" --json
# 4. Combine findings with citations
| User Request | Command | Notes |
|---|---|---|
| "What obligations...?" | search | Semantic search |
| "Are we compliant with...?" | gaps | Gap analysis |
| "Citation for...?" | cite | Citation chain |
| "TPR guidance on...?" | handbook | Handbook search |
| "Related to obligation X" | similar | Semantic similarity |
| "Details of X" | detail | Full enrichment |
Always use --json flag when parsing results programmatically:
import subprocess
import json
result = subprocess.run(
["uv", "run", "apex_intel.py", "search", query, "--json"],
capture_output=True, text=True,
cwd="~/Projects/pensions/apex-governance/tools"
)
data = json.loads(result.stdout)
obligations = data["results"]
Every legal conclusion must be traceable:
Conclusion → Obligation ID → Statutory Provision → Legislation
When reporting findings, always include:
Never state legal positions without citation chains.
This skill uses ~1,200 tokens vs ~8,000+ tokens for full MCP stack.
| Approach | Context Cost | Use Case |
|---|---|---|
| apex-intel CLI | ~1,200 tokens | Targeted queries |
| Full MCP stack | ~8,000+ tokens | Exploratory research |
Recommendation: Use CLI for 80% of queries, MCP for complex multi-hop exploration.
Auto-start (default): The CLI will automatically start the API backend if not running.
Manual requirements:
backend/.venv (created with uv venv --python 3.12 .venv)Optional:
doppler run -- uv run apex_intel.py ...Disable auto-start:
uv run apex_intel.py --no-auto-start search "query"
# or
APEX_AUTO_START=false uv run apex_intel.py search "query"