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 skills add hackersandwizards/artificer --skill improve-workflowgit clone --depth 1 https://github.com/hackersandwizards/artificerWrote 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/skills/hackersandwizards/artificer/improve-workflow)<a href="https://agentmods.dev/skills/hackersandwizards/artificer/improve-workflow"><img src="https://agentmods.dev/badge/skills/hackersandwizards/artificer/improve-workflow.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 5.1 | $0.00085 | $0.00680 |
| Opus 5 | $0.00043 | $0.00340 |
| Sonnet 5 | $0.00017 | $0.00136 |
| Haiku 4.5 | $0.00009 | $0.00068 |
Grade A, and why
improve-workflow 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 8d 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 — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Work in the main context; do not fork. This loop drives the Workflow tool, a Monitor, and task notifications, which only the main loop can orchestrate (a subagent cannot spawn subagents).
Input: $ARGUMENTS is the workflow name followed by a concrete input to run it on.
1. Run
- Start any precondition the workflow assumes (for example
npm run devfor a stage that drives the browser) and confirm it is up before launching. - Launch the workflow with the Workflow tool, passing the concrete input as
args. Keep the transcript dir and runId from the result. - If the workflow launches freshly generated agents by
agentType, reload first. A mid-session agent file does not register:agent({agentType})resolves against the registry frozen at session start.
2. Observe without flooding context
- Arm a tree-change Monitor for a stage-by-stage play-by-play: poll
git status --shortand emit when the changed-file set changes. - Read
<transcript-dir>/journal.jsonl. It maps each agentId to its started and result records and holds each stage's return value, so it is the spine of the run. - For stage detail, grep the per-stage
agent-*.jsonl. Never read one whole: a subagent transcript is the full conversation and overflows context.
3. Verify the assumptions
- Gates fire: grep transcripts for
gate not satisfied. A passing gate is silent, so absence proves nothing. To prove a gate fires, catch a stage that stops while its gate is unsatisfied, or run a throwaway probe agent whose gate appends to a log file (always exit 0 so it cannot loop), and reload so itsagentTyperesolves. - Structured returns thread and branch: read the verdict objects in the journal and confirm the workflow looped or broke on them as intended.
- Parallel stages overlap: in the journal, two stages show
startedbefore either showsresult. - The end state is real: run the check the workflow claims to enforce against the working tree.
4. Diagnose
For anything that did not progress as intended, trace it to the exact stage, agent, or workflow line, reading that one transcript with targeted greps.
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.
- 8d ago First seen · 42 lines · 85 tokens per session scan A 1b52f752e3d7
improve-workflow is a skill published in the GitHub repository hackersandwizards/artificer (16 stars, last pushed 2mo ago), licensed MIT. It adds 85 tokens to every session and 680 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
patpat-verify
Prove an implementation or repository claim against its authoritative artifact or user surface. Use after code changes, during acceptance checks, or whenever tests and builds are insufficient evidence of real behavior.
patpat-eval
Evaluate whether an agent skill triggers and performs correctly using isolated, evidence-based trials. Use after non-trivial skill changes or when routing quality is uncertain.
patpat-verifier
Create or maintain a project-specific verification skill that launches, exercises, observes, and cleans up the real system. Use when a repository lacks repeatable proof of its user-facing behavior.
patpat-perf
Diagnose a resource regression or drive sustained numeric optimization against a measurable target. Use for latency, CPU, memory, throughput, startup, or other performance implementation when comparable before-and-after measurement is practical.
python-sdk
Implement or modify Python SDK behavior under python/composio, including tools, toolkits, sessions, auth configs, connected accounts, client integration, and shared Python models. Use for Python core runtime/API work; pair with python-testing and cross-sdk-parity when TypeScript must match.
typescript-providers
Implement, modify, test, or document TypeScript provider packages under ts/packages/providers, including framework adapters for OpenAI, Anthropic, Google, LangChain, Mastra, Vercel, LlamaIndex, Cloudflare, and Claude Agent SDK. Use for provider-specific TS work; do not use for core-only changes.