semantic

A command-line tool for running UKB knowledge-base workflows and managing their ontology, which is the system used to organize types of entities and their relationships. It sends workflow runs to a local service so the dashboard can display their progress.

In plain words
What is it for?
Running wave-analysis workflows, checking service health, listing available semantic tools, inspecting ontology classes, and refreshing outdated knowledge entities.
Why use it?
It provides one place to start, debug, inspect, and monitor knowledge-base analysis instead of running workflow code in a way the dashboard cannot see. Different commands distinguish real language-model runs from mock debugging runs.

Command for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/fwornle/coding/semantic
Clone the repo
git clone --depth 1 https://github.com/fwornle/coding

Made for: Claude Code.

Per session 76 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 945 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00076 $0.00945
Opus 5 $0.00038 $0.00473
Sonnet 5 $0.00015 $0.00189
Haiku 4.5 $0.00008 $0.00094

Measured yesterday against content hash c2090173f061, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

semantic scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

- **Service health:** `curl -s localhost:3848/health`
.claude/commands/semantic.md · 83 lines

How it starts

The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/semantic

The semantic CLI (bin/semantic) is the entry point to the semantic-analysis tools. It POSTs to /tool/<name> on the SSE server the coding-services container already runs at :3848.

That transport is deliberate, not incidental: execute_workflow dispatches into an in-process workflow state machine, and the broadcaster subscribed to it is what feeds the dashboard's /workflow-events view. Running the same handler in a one-shot process would start the workflow and leave the UI blind to it.

  • Dashboard: http://localhost:3032 → Performance → UKB
  • Service health: curl -s localhost:3848/health
  • Every available tool: semantic tools

🚨 UKB Workflow Control — match the flags to what the user actually said

User says Mode Command
"ukb", "full ukb", "ukb full" PRODUCTION — real LLM calls, runs for 10–20 min semantic workflow run wave-analysis --team coding
"ukb full debug", "ukb debug" DEBUG — mock LLM, single-step, step-into-substeps semantic workflow run wave-analysis --team coding --debug
  • NEVER default to debug unless the user explicitly said "debug".
  • NEVER run bare ukb as a shell command — that is the retired legacy script, not this CLI.
  • Long workflows are async by default: the command returns a workflow id and leaves the run going. Watch it on the dashboard or poll semantic workflow status. Do not wait on it.

Workflow names: wave-analysis (the UKB pass), batch-analysis, incremental-analysis, complete-analysis.

Commands

semantic workflow run wave-analysis --team coding          # production UKB pass
semantic workflow run wave-analysis --team coding --debug  # mock LLM + single-step
semantic workflow run batch-analysis --param fullAnalysis=true
semantic workflow status                                   # progress of the current run

semantic ontology status --team coding      # config, paths, validation mode, registered teams
semantic ontology classes --team coding     # available entity classes
semantic ontology suggest --team coding     # propose new classes from unclassified entities
semantic ontology inject --team coding      # hot-swap an ontology config

semantic refresh KnowledgePersistencePattern --team coding
semantic refresh '*' --team coding --dry-run          # preview a batch refresh
semantic refresh '*' --team coding --parallel 8

semantic checkpoint reset --days-ago 7      # re-analyse from N days back

semantic insights notes.md --type architecture
semantic analyze code src/foo.ts --focus security
semantic analyze repo /Users/Q284340/Agentic/coding --max-files 200

semantic tools                              # names + descriptions of all 19 tools
semantic tool <name> '{"key":"value"}'      # generic escape hatch for anything unmapped

Read the full file on GitHub · 83 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 83 lines · 76 tokens per session scan A c2090173f061

Subscribe to this mod's changes

semantic is a command published in the GitHub repository fwornle/coding (2 stars, last pushed yesterday), licensed MIT. It adds 76 tokens to every session and 945 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.