Launch week
Security, performance and a rollback plan before real users arrive.
4 of 4 available · 0 paused
A pre-launch checklist for an AI-built app
Fieldbook’s Launch week kit helps you choose agent skills for a security review, performance investigation or release check. Start with the gap in your evidence. Installing every skill does not make an app production-ready, and a passing local check does not authorize deployment.
Choose the check you need
- Unknown exposure: start with an application-security review. Name the login, authorization, input and dependency boundaries that matter. Use the security skill guide to scope permitted tests before picking a skill.
- A slow page: measure a reproducible page and journey before changing code. Record the device, network conditions and metric, then compare under the same conditions. A performance skill needs the tools and access stated in its listing.
- An unproven release path: inspect build checks, environment requirements and the recovery procedure. Distinguish a code rollback from reversing a data migration; restoring an older build may not restore changed data.
- Uncertain behavior: use the testing guide for the critical journey, then review the final patch. Keep failed, skipped and blocked checks visible.
The listings below describe each member’s scope and requested access. Read its review and source before installing one. Paused members remain visible for context but are excluded from install commands. A retained scan is not a check of today’s upstream instructions.
Write a bounded launch brief
Replace the bracketed details before handing this to your agent. This is an illustrative brief, not evidence of a completed audit or a claim that a particular skill will cover every risk.
Review [repository and revision] for [critical user journey] before launch. Begin read-only. Identify the project’s build and test commands and propose the checks needed for this journey. Permitted test targets: [local or staging URL and synthetic accounts]. Out of scope: [systems and data]. Ask before running commands that send data or change state. Separate confirmed defects from risks you could not test. For each finding, give the relevant path, evidence, impact and smallest proposed fix. Do not deploy, run migrations, change production settings, or contact third-party systems.
For example, an app that saves private notes needs evidence that account B cannot read account A’s notes. A successful login test only checks login. Use synthetic notes and test accounts in an authorized environment, and test the server’s permission check rather than only hiding a button.
Collect evidence before deciding to ship
- Revision: the exact commit or working tree checked, plus the environment and relevant configuration without secrets.
- Critical journey: what was attempted, expected and observed, including one meaningful failure or permission boundary.
- Commands: the actual build and test results from the final code, not a list of commands the agent intends to run.
- Recovery: who can revert a release, what happens to stored data and whether the recovery procedure was exercised. Label an untested plan as untested.
- Decision: unresolved issues, their impact, the person accepting any remaining risk and the separate approval to deploy.
When to stop the launch
Stop when a required check fails, a private-data boundary is untested, a secret appears in output or the recovery plan cannot account for a destructive change. Investigate and rerun the affected checks after a fix. A scanner result, an install count or a skill’s review label cannot decide these questions for your project.
Fieldbook’s automated scan covers SKILL.md, not every referenced file or runtime behavior. Read what the review does and does not establish.
Install the available skills
Read each skill’s review and access labels below before installing.
Commands target the repository commits recorded with the automated scans. Scans cover only SKILL.md, not referenced files or runtime behavior. What the review covers.
npx skills@latest add https://github.com/AaravKashyap12/secure-launch/tree/389610e3efc6be85a34555110c3ce69e9195a699 --skill secure-launch
npx skills@latest add https://github.com/addyosmani/agent-skills/tree/39b723ceebf81474d0557c1fecd01eddab585563 --skill shipping-and-launch
npx skills@latest add https://github.com/cloudflare/skills/tree/cbf893b954222276bf1af57e5f23379231e0efb9 --skill web-perf
npx skills@latest add https://github.com/addyosmani/agent-skills/tree/39b723ceebf81474d0557c1fecd01eddab585563 --skill ci-cd-and-automationEach command asks which agents to install for; add -g to install globally. Confirm installation in your chosen agent.
In this kit
Secure Launch
A security checklist shaped to your project before it goes live, with approved fixes checked one by one.
What it can touch: Edits code · Runs commands
Automated scan recorded · evidence and limits
- Automated scan: Checked 16 Nov 2026. Checks SKILL.md patterns and frontmatter; does not run the skill.
- Source revision: 389610e3efc6 (opens in a new tab). The install command targets this commit. Only SKILL.md was scanned.
- Human reading: Full reading recorded 27 Sept 2026; not bound to the current source revision.
- Runtime evidence: Published evaluation notes (opens in a new tab). Evaluation date not recorded here; see the notes for dates, scope and limitations. Not a test of every upstream revision.
A review is not a security audit or a guarantee.
Commands target the repository commits recorded with the automated scans. Scans cover only SKILL.md, not referenced files or runtime behavior. What the review covers.
1installs0 today
npx skills@latest add https://github.com/AaravKashyap12/secure-launch/tree/389610e3efc6be85a34555110c3ce69e9195a699 --skill secure-launchShipping and launch
A pre-launch checklist with feature flags, staged rollout and a rollback plan.
What it can touch: Writes docs
Automated scan recorded · evidence and limits
- Automated scan: Checked 16 Nov 2026. Checks SKILL.md patterns and frontmatter; does not run the skill.
- Source revision: 39b723ceebf8 (opens in a new tab). The install command targets this commit. Only SKILL.md was scanned.
- Human reading: No full human-reading date recorded.
- Runtime evidence: No runtime-test evidence recorded here.
A review is not a security audit or a guarantee.
Commands target the repository commits recorded with the automated scans. Scans cover only SKILL.md, not referenced files or runtime behavior. What the review covers.
71Kinstalls+552 today
npx skills@latest add https://github.com/addyosmani/agent-skills/tree/39b723ceebf81474d0557c1fecd01eddab585563 --skill shipping-and-launchWeb performance
Audits Core Web Vitals from a real performance trace, then maps fixes to your code.
What it can touch: Browser · Needs MCP
Automated scan recorded · evidence and limits
- Automated scan: Checked 16 Nov 2026. Checks SKILL.md patterns and frontmatter; does not run the skill.
- Source revision: cbf893b95422 (opens in a new tab). The install command targets this commit. Only SKILL.md was scanned.
- Human reading: No full human-reading date recorded.
- Runtime evidence: No runtime-test evidence recorded here.
A review is not a security audit or a guarantee.
Commands target the repository commits recorded with the automated scans. Scans cover only SKILL.md, not referenced files or runtime behavior. What the review covers.
129Kinstalls+830 today
npx skills@latest add https://github.com/cloudflare/skills/tree/cbf893b954222276bf1af57e5f23379231e0efb9 --skill web-perfCI/CD and automation
Sets up CI quality gates and deployment strategies in GitHub Actions.
What it can touch: Edits code · Runs commands
Automated scan recorded · evidence and limits
- Automated scan: Checked 16 Nov 2026. Checks SKILL.md patterns and frontmatter; does not run the skill.
- Source revision: 39b723ceebf8 (opens in a new tab). The install command targets this commit. Only SKILL.md was scanned.
- Human reading: No full human-reading date recorded.
- Runtime evidence: No runtime-test evidence recorded here.
A review is not a security audit or a guarantee.
Commands target the repository commits recorded with the automated scans. Scans cover only SKILL.md, not referenced files or runtime behavior. What the review covers.
71Kinstalls+544 today
npx skills@latest add https://github.com/addyosmani/agent-skills/tree/39b723ceebf81474d0557c1fecd01eddab585563 --skill ci-cd-and-automation