Install
$ agentstack add skill-leochens-skill-pre-open-source-skill-pre-open-source ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ● Environment & secrets Used
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Pre-Open Source
Goal
Prepare a project for public release by scanning the repository, identifying risks, recommending fixes, generating the minimum necessary open-source documents/configuration, and optionally applying safe changes after confirmation.
Operating Rules
- Treat open-sourcing as a security and product-readiness task, not only documentation cleanup.
- Start read-only when the user asks for an assessment, audit, review, or "看看能不能开源".
- Do not delete files, rewrite Git history, publish, push, create releases, or expose suspected secrets without explicit user confirmation.
- Do not print secrets in chat. Redact tokens, keys, private URLs, emails, and credentials in findings.
- Prefer a clean migration to a new public repository when history may contain secrets, proprietary files, or messy private work.
- State that license guidance is practical engineering guidance, not legal advice.
Workflow
- Scan the project
- Inspect repository structure, key entrypoints, package/build files, docs, config files,
.gitignore, templates, CI files, license files, and Git state. - Check whether the project appears to contain private assets, environment files, generated outputs, build artifacts, logs, datasets, credentials, or internal documentation.
- If Git history is relevant, inspect recent commits and known risky paths without dumping sensitive content.
- Clarify intent
- Ask only for missing high-impact context: project name, one-sentence description, target audience, intended license family, public/private boundaries, and whether edits should be applied automatically.
- If the user already supplied enough context, proceed and mark unknowns as assumptions.
- Assess risks
- Report risks by severity: blockers, should-fix-before-release, nice-to-have.
- Cover sensitive content, license ambiguity, missing docs, build/test uncertainty, dependency/security concerns, and repository hygiene.
- Use
references/checklist.mdwhen a full readiness pass is requested.
- Confirm the plan
- Before changing files, summarize the proposed generated/modified files and any destructive or irreversible action.
- For safe documentation/config additions, proceed when the user asked you to prepare or implement.
- For history rewriting, deleting files, publishing, pushing, or public release setup, require explicit confirmation.
- Generate and apply
- Generate only the files that match the project and user intent.
- Use existing project tone, language, package manager, test commands, and framework conventions.
- Prefer concrete placeholders such as `` only when the value is genuinely unknown.
- Keep docs honest: do not claim tests, CI, security scans, API stability, or compatibility that was not verified.
- Final check
- Re-scan changed files for accidental sensitive content.
- Run focused formatting, docs lint, build, or test commands when available and proportionate.
- Deliver a concise readiness summary with remaining blockers and exact next steps.
Default Deliverables
Generate or improve these when appropriate:
README.md: project intro, features, quick start, usage, configuration, examples, FAQ/troubleshooting, contribution note, license note.LICENSE: choose MIT, Apache-2.0, BSD-3-Clause, GPL-3.0, AGPL-3.0, or a user-specified license; explain tradeoffs briefly before applying..gitignore: cover language/framework/tooling, dependency directories, build outputs, logs, IDE files, OS files, temp files, local env files, and large generated artifacts.CONTRIBUTING.md: setup, branch/commit expectations, tests, issue/PR flow, code style, release process.CODE_OF_CONDUCT.md: include only if requested or suitable for a community-facing project; use Contributor Covenant as a common default..github/ISSUE_TEMPLATE/*and.github/PULL_REQUEST_TEMPLATE.md: bug report, feature request, and PR checklist.CHANGELOG.md: use Keep a Changelog style when the project has versions or release notes.SECURITY.md: vulnerability reporting path and supported versions when the project expects external users.- CI config: add minimal build/test/lint workflows only when requested and compatible with the repo.
- Readiness report/checklist: summarize what is ready, what changed, what remains risky, and what should be verified before publishing.
Sensitive Content Strategy
When sensitive content is suspected:
- Stop broad generation work until the risk is triaged.
- Identify the file/path/category and reason without exposing the secret value.
- Recommend immediate rotation/revocation for exposed keys, tokens, passwords, certificates, or private endpoints.
- Move required configuration into environment variables and provide
.env.examplewith comments. - Add or strengthen
.gitignoreso local secrets, logs, data dumps, build outputs, and private assets are excluded. - If history contains sensitive content, recommend the safer clean-repo migration path first:
- Create a new private clean repository.
- Copy only safe source files and generated public docs.
- Recreate needed commits from the clean tree.
- Validate again before making public.
- Use destructive history rewrite tools such as
git filter-repoor BFG only after explicit confirmation, a backup plan, and clear communication that clones/forks/tags may still contain old data.
License Selection Guidance
- MIT: simple permissive default for broad reuse.
- Apache-2.0: permissive with explicit patent grant; often suitable for companies and SDKs.
- BSD-3-Clause: permissive with attribution and non-endorsement language.
- GPL-3.0: copyleft for projects that require derivative works to remain open.
- AGPL-3.0: stronger copyleft for network/server software.
- Proprietary or source-available: use only if the user does not actually want open-source terms.
Always ask for confirmation before writing a license if ownership, employer rights, third-party code, or intended commercialization is unclear.
User Experience
- Present readiness as a checklist with status labels:
Ready,Needs work,Blocked,Unknown. - Keep recommendations actionable: include file paths, reasons, and the smallest next step.
- Separate "can publish now" from "should improve soon".
- If asked to auto-execute, make safe doc/config changes directly, then summarize exactly what changed.
Reference
- For the complete checklist derived from the pre-open-source preparation flow, read
references/checklist.md.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Leochens
- Source: Leochens/skill-pre-open-source
- License: MIT
- Homepage: https://world.guantou.site/
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.