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 Terryc21/workflow-audit --skill workflow-auditgit clone --depth 1 https://github.com/Terryc21/workflow-auditWrote 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/terryc21/workflow-audit/workflow-audit)<a href="https://agentmods.dev/skills/terryc21/workflow-audit/workflow-audit"><img src="https://agentmods.dev/badge/skills/terryc21/workflow-audit/workflow-audit/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/terryc21/workflow-audit/workflow-audit"><img src="https://agentmods.dev/badge/skills/terryc21/workflow-audit/workflow-audit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 4 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high System Prompt Leakage · line 165 Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.Fix: Remove any instructions that reveal, print, or output system prompts or internal rules. System instructions should never be exposed to end users.
- high System Prompt Leakage · line 171 Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.Fix: Remove any instructions that reveal, print, or output system prompts or internal rules. System instructions should never be exposed to end users.
- medium Excessive Agency · line 81 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 100 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00059 | $0.06623 |
| Opus 5.5 | $0.00024 | $0.02649 |
| Sonnet 5.5 | $0.00012 | $0.01325 |
| Haiku 4.5 | $0.00006 | $0.00662 |
Grade A, and why
workflow-audit 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 26d 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 — 482 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Workflow Audit Skill
Quick Ref: 5-layer UI workflow audit: discover entry points → trace flows → detect issues → evaluate UX → verify data wiring. Output:
.workflow-audit/in project root.
You are performing a systematic workflow audit on this SwiftUI application.
Rating findings. Where a layer produces findings — a confirmed problem you have read the code for — rate each one on all six dimensions in the Issue Rating Table: Urgency, Risk:Fix, Risk:NoFix, ROI, Blast Radius, and Fix Effort. A partial rating quietly drops a finding out of triage: the user is deciding what to fix before the next release, and a row missing Risk:NoFix or Blast Radius can't be weighed against the rows that have them, however real the bug is.
Layer 1 is the exception, by design. Its output is an inventory of entry points with flags for Layer 2 to investigate — hypotheses, not verified problems. Rating an unverified flag would assign urgency to something nobody has confirmed is broken, which is exactly what the Work Receipts rule exists to prevent. Layer 1 reports counts and flags; ratings begin at Layer 3 where findings are verified against source.
Quick Commands
| Command | Description |
|---|---|
/workflow-audit |
Full 5-layer audit |
/workflow-audit layer1 |
Discovery only — find all entry points |
/workflow-audit layer2 |
Trace — trace critical paths |
/workflow-audit layer3 |
Issues — detect problems across codebase |
/workflow-audit layer4 |
Evaluate — assess user impact |
/workflow-audit layer5 |
Data wiring — verify real data usage |
/workflow-audit trace "A → B → C" |
Trace a specific user flow path |
/workflow-audit diff |
Compare current findings against previous audit |
/workflow-audit fix |
Generate fixes for found issues |
/workflow-audit fix --apply-anyway |
Same, but auto-apply even with uncommitted changes |
/workflow-audit status |
Show audit progress and remaining issues |
What ships with it
27 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.
- agents-skill/examples/bad-patterns.swift 11 KB
- agents-skill/examples/good-patterns.swift 7.8 KB
- agents-skill/issue-categories.md 14 KB
- agents-skill/pattern-library.md 11 KB
- agents-skill/quick-start.md 5.1 KB
- agents-skill/README.md 8.0 KB
- agents-skill/templates/layer1-inventory.yaml 2.2 KB
- agents-skill/templates/layer2-flow-trace.yaml 2.6 KB
- agents-skill/templates/layer3-issue.yaml 3.8 KB
- agents-skill/templates/layer4-evaluation.yaml 5.1 KB
- agents-skill/templates/layer5-data-wiring.yaml 5.8 KB
- agents/layer1-patterns.md 4.7 KB
- agents/layer2-methodology.md 5.9 KB
- agents/layer3-issue-detection.md 55 KB
- agents/layer4-semantic-evaluation.md 8.4 KB
- agents/layer5-data-wiring.md 10 KB
- agents/README.md 10 KB
- examples/2026-04-15-workflow-audit-stuffolio.md 14 KB
- examples/sample-stuffolio-scan/README.md 1.5 KB
- examples/sample-stuffolio-scan/sample-layer1-inventory.yaml 10.0 KB
- examples/sample-stuffolio-scan/sample-layer1-summary.md 3.9 KB
- examples/sample-stuffolio-scan/sample-layer2-summary.md 4.3 KB
- examples/sample-stuffolio-scan/sample-layer2-traces/flow-001-pricewatch.yaml 3.6 KB
- examples/sample-stuffolio-scan/sample-layer2-traces/flow-002-repairadvisor.yaml 4.7 KB
- examples/sample-stuffolio-scan/sample-layer2-traces/flow-003-bulkactions.yaml 4.7 KB
- examples/sample-stuffolio-scan/sample-layer3-results.yaml 8.2 KB
- radar-suite-core.md 57 KB
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.
- 26d ago Changed · +1 lines bd5a94ef7669
- 1mo ago First seen · 481 lines · 59 tokens per session scan A 6b353eefd600
workflow-audit is a skill published in the GitHub repository Terryc21/workflow-audit (58 stars, last pushed 27d ago), licensed Apache-2.0. It adds 59 tokens to every session and 6,623 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 2026-08-30.
Other skills, from other repositories
ui-enhancer-radar
Systematic iOS/SwiftUI UI audit with design intent interview, 14-domain analysis (including Color Audit with adaptive Color Profile, iPad Sheet Sizing caller-side audit, Button Hit Region three-factor interaction audit, and Silent Picker menu-presentation audit), element compaction, cross-view consistency checks…
ui-path-radar
UI path tracer for SwiftUI/UIKit apps. 5-layer audit with 34 issue categories and 19 automated checks: discover entry points, trace flows, detect dead ends and broken promises, evaluate UX impact, verify data wiring. Supports targeted trace, diff against previous audits, and handoff to planning skills. Triggers…
roundtrip-radar
Per-journey code audit tracing data through complete user flows for bugs, data safety, performance, and round-trip completeness. Discovers workflows, audits each end-to-end, rolls up cross-cutting issues, and supports natural-language flow tracing. Triggers: "roundtrip audit", "trace user journey", "/roundtrip-radar".
capstone-radar
Unified A-F grading and ship/no-ship decisions for the 6-skill radar family (5 companions + capstone). Aggregates handoffs from data-model, ui-path, roundtrip, time-bomb, and ui-enhancer; owns 5 grep-reliable domains; tracks velocity; celebrates improvements. Triggers: "capstone radar", "can I ship", "grade codebase"…
time-bomb-radar
Finds deferred operations that crash on aged data -- code that passes every test but breaks weeks or months after release. Covers cascade deletes, cache expiry, trial paths, background accumulation, date-threshold transitions, and scheduled side effects. Triggers: "time bomb", "time-bomb", "/time-bomb-radar", "aged…
radar-suite-axis-classification
Shared axis classification framework for all radar-suite skills. Every finding must be classified as axis1 (bug), axis2 (scatter), or axis3 (dead/smelly) before emission, with mandatory coaching fields and file:line citations to existing patterns in the audited codebase. Triggers: invoked by every radar before…