Small, testable automation work for spreadsheets, CI pipelines, and AI coding agents.
Every engagement starts with a written scope and acceptance cases. I do not claim revenue, awards, or customer results that have not been independently verified.
Need a quick fit and price? See paid support and delivery. The request forms are inquiry-only; a supported payment route is confirmed before any payment is requested.
Use one of the public samples before requesting paid work:
- Spreadsheet preflight checklist
- GitHub Actions: the first 15 minutes
- AI agent skill acceptance checklist
- Go cross-architecture workflow
- Korean AGENTS.md guide and examples
Browse all samples and matching service scopes
- Ten yes/no checks for originals, formulas, hidden structures, external links, units, overrides, control totals, and independent review
- Runs entirely in the browser without uploads, accounts, or stored answers
- Gives a score-specific path to the free checklist, spreadsheet audit, USD 39 ebook + CLI, or the Developer Reliability Bundle
- Start here: run the free scorecard
- Ten yes/no checks for reproducibility, permissions, secrets, cache, logs, ownership, and runtime coverage
- Runs entirely in the browser without uploads, accounts, or stored answers
- Gives a score-specific path to free material, CI diagnosis, or the Developer Reliability Bundle
- Start here: run the free scorecard
- Launch price: USD 49
- Format: five independently versioned ZIP products in one verified bundle
- Includes: two Korean technical ebooks, three execution kits, checksums, and a delivery checklist
- Product details: review the full bundle
The bundle combines spreadsheet preflight, GitHub Actions triage, AGENTS.md auditing, OSS bounty monitoring, and Go/Linux cross-architecture CI. Every internal ZIP passes its own archive check, and the outer package rebuild is deterministic. The full paid bundle is not published in this repository.
Ask about the reliability bundle
- Launch price: USD 39
- Format: 23-page Korean PDF
- Extras: one-page checklist, dependency-free read-only Python CLI, and quick start
- Free preview: read the first 6 pages
- Product details: review the full scope
The guide explains seven structural workbook signals before an XLSX/XLSM file is opened for editing. The CLI reports those signals as JSON without modifying the workbook. It does not calculate formulas, analyze VBA, inspect external targets, or certify that a file is safe. The full paid ZIP is not published in this repository.
Ask about the spreadsheet ebook
- Launch price: USD 39
- Format: 19-page Korean PDF
- Extras: incident report template and 12-item triage checklist
- Free preview: read the first 6 pages
- Product details: review the full scope
The ebook classifies failures into code, workflow or runner, account-owner, and external-state paths before a patch is attempted. The full paid ZIP is not published in this repository.
Free troubleshooting guide: prove whether GitHub Actions exit code 137 is OOM.
- Price: USD 19
- Format: 16-page Korean PDF
- Extras: operations checklist and persistence audit template
- Free preview: read the first 6 pages
- Product details: review the full scope
The handbook covers instruction hierarchy, loading limits, fresh-process persistence checks, fail-closed external-model gates, and background pipeline recovery. The full paid ZIP is not published in this repository.
Free troubleshooting guide: diagnose Codex AGENTS.md loading, overrides, and truncation.
- Launch price: USD 39
- Targets: Linux AMD64, ARM64, and 386 with
CGO_ENABLED=0 - Extras: local compile report, optional QEMU runtime smoke, Dockerfile, working example, and adoption checklist
- Free diagnostic CLI: inspect an
exec format errorlocally - Free troubleshooting guide: diagnose Docker AMD64 vs ARM64 mismatches
- Free preview: read the actual compile workflow
- Product details: review the full scope
The kit passed ten self-tests and records each target result in a TSV report. It does not include a CGO cross-toolchain, private-repository installation, paid runner capacity, or a guarantee that every platform will build.
- Launch price: USD 39
- Runtime: dependency-free, read-only Python 3.13+ CLI
- Checks: instruction discovery, required sections, backtick paths, explicit conflicts, and the default 32 KiB file and effective-chain boundary
- Extras: six role and nested-scope templates, a 15-item adoption checklist, and a Korean team worksheet
- Free guide: read the guide and three examples
- Product details: review the full scope
The tool produces human-readable and JSON reports without executing repository commands, modifying files, using telemetry, or uploading repository content. Static checks do not guarantee agent performance or universal compatibility. The full paid ZIP and auditor source are not published in this repository.
| Package | Price | Delivery |
|---|---|---|
| Basic | KRW 49,000 | Structure report for one XLSX/XLSM file |
| Standard | KRW 99,000 | Report, risk ranking, and remediation checklist |
| Advanced | KRW 149,000 | Standard package plus a tested repair plan |
The audit can inventory formulas, cached error values, hidden sheets, merged ranges, external links, macros, and calculation settings without modifying the source workbook.
| Package | Price | Delivery |
|---|---|---|
| Diagnosis | KRW 79,000 | Root-cause report and reproducible verification |
| Patch | KRW 149,000 | Diagnosis plus a minimal tested patch |
| Complex | From KRW 299,000 | Multi-platform or multi-workflow repair |
The repository and failing run must be shareable before work begins. Account, secret, billing, runner-capacity, and vendor-outage problems are reported separately from code-fixable failures.
| Package | Price | Delivery |
|---|---|---|
| Review | KRW 79,000 | Route, key-exposure, abuse, and provider-cost risk report |
| Hardening patch | KRW 149,000 | Minimal tested server-side protection patch |
| Multi-route or app scope | From KRW 299,000 | Auth, rate limits, redaction, tests, and operator notes |
The work can cover server-only provider keys, authentication, bounded rate limits, safe 401/429 responses, cost controls, redacted errors, and offline tests. No credentials or production access are required for the initial fit check.
For a bounded first engagement, the AI API cost/security diagnosis is KRW 79,000. It covers up to two public API routes and delivers a reproducible risk report with pass/fail acceptance checks. Read the synthetic sample report first, then use the login-free public catalog contact path with a repository or minimal reproduction. The GitHub request form is an optional logged-in route. Do not send keys or private customer data. Scope and payment method are agreed in writing before work begins.
| Package | Price | Delivery |
|---|---|---|
| Basic | KRW 99,000 | One focused skill with examples and checks |
| Standard | KRW 299,000 | Skill, tool integration, fixtures, and evaluation |
| Custom | Quoted | Multi-tool or repository-integrated agent workflow |
The deliverable is a reusable instruction package with explicit inputs, outputs, refusal boundaries, failure examples, and repeatable validation—not an untested prompt.
| Product | Price |
|---|---|
| Developer Reliability Bundle, 5 products | USD 49 |
| Spreadsheet Preflight Ebook + CLI | USD 39 |
| AGENTS.md Audit Kit + templates | USD 39 |
| GitHub Actions CI Triage Ebook + templates | USD 39 |
| Open Source Bounty Monitor Kit | USD 19 |
| Go/Linux Cross-Architecture CI Starter Kit | USD 39 |
GitHub records purchase interest only; it does not process payment. A supported payment route and delivery method are confirmed before any payment is requested.
Public work currently includes:
- 20 open upstream pull requests monitored with a read-only operations loop;
- tested fixes across C, C++, Go, shell, and CI workflows;
- a 23-page Korean spreadsheet preflight guide and workbook structure auditor;
- a 19-page GitHub Actions failure triage guide;
- verified Linux AMD64, ARM64, and 386 cross-architecture build examples.
- a read-only AGENTS.md auditor with nine tests and six installable templates.
Pull requests are evidence of implementation work, not proof of reward, acceptance, or revenue.
- Open the matching request form without attaching secrets or private data.
- Receive a written scope, exclusions, price, and acceptance cases.
- Provide only the minimum sanitized input needed for the agreed scope.
- Receive the artifact, validation commands, and known limitations.
No work starts and no payment is requested until both sides have a written scope. Public issues must never contain credentials, customer data, private source code, financial records, or proprietary workbooks.