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 yunzeforbetter/CastFlow --skill skill-doctorgit clone --depth 1 https://github.com/yunzeforbetter/CastFlowWrote 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/yunzeforbetter/castflow/skill-doctor)<a href="https://agentmods.dev/skills/yunzeforbetter/castflow/skill-doctor"><img src="https://agentmods.dev/badge/skills/yunzeforbetter/castflow/skill-doctor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yunzeforbetter/castflow/skill-doctor"><img src="https://agentmods.dev/badge/skills/yunzeforbetter/castflow/skill-doctor.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00050 | $0.00901 |
| Opus 5.5 | $0.00020 | $0.00360 |
| Sonnet 5.5 | $0.00010 | $0.00180 |
| Haiku 4.5 | $0.00005 | $0.00090 |
Grade A, and why
skill-doctor 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 15d 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
skill-doctor
Finish "maintain an existing skill" in one pass. The goal is a skill with a provable gain on this project and the current model. The result may be a fix, a shrink, a retirement, a deletion, or no change.
Yield
Creating a skill, or writing a programmer skill from the generation queue, yields to skill-creator. Compiling a requirement into a package and then stopping yields to goal-loop-creator. This skill runs the optimization itself.
Duties
If the user did not name a target skill, ask. Do not scan the whole repo. Optimize one skill at a time. If ITERATION_GUIDE.md is missing, stop. Do not invent a focus.
- Read the target skill's
SKILL.mdandITERATION_GUIDE.md. Turn the guide's triggers into this pass's scope. If the scope is unclear, stop. - Read only the scripts the target skill names, and the definitions this pass actually opened on the call chain. Do not edit business scripts, prefabs, assets, or adapter mirrors.
- Classify health first: locator ok, file missing, feature moved, call chain broken, rule stale, generic content with no gain, or skill harmful. A definition that was not opened is unverified. If the project was not run, do not claim a runtime effect.
- Record the gap, or confirm there is none. A new feature may come only from a script opened this pass, and only if the target skill does not already record it. Do not invent a feature. Do not widen the module scope.
- Freeze the eval prompts before any candidate body exists. Prompts state the finished result, the constraints, and the boundaries. They do not quote the candidate's wording. Put the temporary snapshot, prompts, candidate, and results in
.castflow-runtime/tmp/skill-doctor/<target-skill-name>/. Do not write a maintenance queue, a hash, or Git-tracked state. - Run three groups with the same current model and the same context: no skill, old skill, and candidate skill. Cover at least one target task and one adjacent or out-of-scope task. If a call chain is broken, add a moved or missing scenario. Without an actual run, do not claim the eval passed.
- The candidate writes only a fix that has evidence. Update a stale locator only after the new file is opened and confirmed. An uncertain chain stays unverified. A generic skill that does not beat the current model shrinks or retires.
- Write back to the target skill only when the candidate has a citable advantage over the old skill on the target result, and regression, boundaries, and out-of-scope behavior are not worse. If the old skill is harmful, restoring the no-skill baseline and removing the harm also counts as an improvement. No change, a downgrade, a mixed result, or no difference: do not overwrite.
- Run
py .castflow-runtime/manager.py validatebefore and after the overwrite. After it passes, runpy .castflow-runtime/manager.py sync. On failure, leave the original skill unchanged. Delete the temporary directory when the task ends, unless the user asks to keep the report.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 15d ago Changed · +5 lines · +3 tokens per session 27b525643efd
- 28d ago First seen · 36 lines · 47 tokens per session scan A 59ade82d3a08
skill-doctor is a skill published in the GitHub repository yunzeforbetter/CastFlow (108 stars, last pushed 16d ago), licensed MIT. It adds 50 tokens to every session and 901 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-30.
Other skills, from other repositories
debug-optimize-lcp
Guides debugging and optimizing Largest Contentful Paint (LCP) using Chrome DevTools MCP tools. Use this skill whenever the user asks about LCP performance, slow page loads, Core Web Vitals optimization, or wants to understand why their page's main content takes too long to appear. Also use when the user mentions…
systematic-debugging
Use when debugging a failing test, build error, or runtime issue that isn't immediately obvious. Guides a 4-phase root cause analysis instead of random fix attempts.
log-error-digest
Analyze log files to troubleshoot errors, identify peak error periods, and produce error clustering, frequency statistics, and time distribution reports. Supports JSON, syslog, and Nginx formats with automatic detection. Use when a user uploads a .log file and asks to analyze errors, find patterns, debug issues, or…
cwv-optimizer
Diagnose and fix Core Web Vitals issues on AEM Edge Delivery Services pages. Goes deeper than generic CWV advice by understanding EDS-specific performance patterns including the 100KB LCP budget, E-L-D loading phases, block rendering behavior, and third-party script impact. Produces specific fixes for LCP, CLS, and…
incident-response-incident-response
Coordinate production incident triage, mitigation, investigation, recovery, and follow-up. Use when responding to outages or other incidents that require structured operational coordination.
ci-fixer
CI failures - read error, minimal fix, verify.