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.
npx agentmods add commands/hesreallyhim/ultracode-workflows/wf-watchgit clone --depth 1 https://github.com/hesreallyhim/ultracode-workflowsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/hesreallyhim/ultracode-workflows/wf-watch)<a href="https://agentmods.dev/commands/hesreallyhim/ultracode-workflows/wf-watch"><img src="https://agentmods.dev/badge/commands/hesreallyhim/ultracode-workflows/wf-watch.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00029 | $0.00848 |
| Opus 5 | $0.00015 | $0.00424 |
| Sonnet 5 | $0.00006 | $0.00170 |
| Haiku 4.5 | $0.00003 | $0.00085 |
Grade A, and why
wf-watch scanned grade A with 0 findings 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 3d ago.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/wf-watch
Watch one workflow run in a focused visual window: the run's drawn topology — a rail with one band per phase, a lane per agent fanning out and joining at parallel() barriers, pipeline stages flowing without a join wall, panels/child workflows as containment, loops as return arcs — with per-agent label, duration, tokens and model on the lanes themselves, and the longest/heaviest agents quietly emphasised. It writes a single self-contained HTML file (no server, no network, no JavaScript — the page carries a script-src 'none' CSP) and opens it from file://. It renders for both light and dark system themes, and every degenerate case (script not yet persisted, analyzer refusal, empty journal, killed run) gets a designed state that says what is known and why the rest is not.
It sits alongside the lower-fidelity statusline, which shows observed activity as a single ambient line without reading workflow source.
Steps
If the user wants to see a workflow before running it, use --script with the script's path: the topology renders pending without executing the workflow body. Describe statically known multiplicities precisely: exact counts are exact, bounded or conditional fan-outs are ranges, and runtime-data-dependent fan-outs are unknown. Do not turn a range or unknown width into a definite spend estimate. If the user is deciding whether to trust the workflow, also send them to /wf-explain; topology alone is not a safety review.
Resolve the target from $ARGUMENTS: a run id if given, otherwise the latest run (which prefers one still in flight). If the user asked for live/watch behavior, include --watch.
# BEFORE running: the plan alone — shape and agent count, nothing live
node "${CLAUDE_PLUGIN_ROOT}/scripts/watch-run.mjs" --script <path-to-workflow.js> --open
# snapshot of the latest run
node "${CLAUDE_PLUGIN_ROOT}/scripts/watch-run.mjs" --latest --open
# a specific run
node "${CLAUDE_PLUGIN_ROOT}/scripts/watch-run.mjs" --run <run-id> --open
# live updates while a run is in flight (regenerates every 2s until it finishes)
node "${CLAUDE_PLUGIN_ROOT}/scripts/watch-run.mjs" --latest --open --watch
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.
- 3d ago First seen · 44 lines · 29 tokens per session scan A dbaecffb9ec7
wf-watch is a command published in the GitHub repository hesreallyhim/ultracode-workflows (2 stars, last pushed 27d ago), licensed MIT. It adds 29 tokens to every session and 848 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
export-md
Export this session's conversation to a markdown file.
health
Check running Claude Code or Codex sessions, find zombies, offer to clean up safely.
quick
Quick 10-second context health check with quality score and top issues.
no-cape
Manually load a no-cape discipline skill and apply it to the current work.
review
Cold re-quiz on code that already shipped — your own session commits, not the change in front of you.
delegate
Delegate a coding task to the Cursor CLI agent (Composer by default).