AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Oss Bootstrap

skill-olgaiv39-claude-oss-skills-oss-bootstrap · by olgaiv39

Prepare a new public open-source repository, or convert a small existing directory into a clean publishable starting point. Use for an empty or nearly empty repo, a first local commit, or a first remote push. Do not use for routine feature work or releasing an established project. Never commits, creates a remote, or pushes unless explicitly requested.

— No reviews yet
0 installs
27 views
0.0% view→install

Install

$ agentstack add skill-olgaiv39-claude-oss-skills-oss-bootstrap

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No 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.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-olgaiv39-claude-oss-skills-oss-bootstrap)

Reliability & compatibility

✓ Security review passed
0 installs to date
— no reviews yet
● 2mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Oss Bootstrap? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

oss-bootstrap

Prepare the smallest publishable repository from a new or nearly empty directory. Inspect before writing. Never overwrite existing README, LICENSE, or .gitignore, and never commit, add a remote, or push without an explicit request.

Activate when

  • Starting a new open-source project
  • The repository is empty or nearly empty
  • Converting a private prototype into a clean public starting point
  • Preparing the first local commit or first remote push

Do not activate when

  • Doing routine feature work in an established repository -> use implement-minimal
  • Releasing or deploying an existing project -> use release-deploy
  • Migrating a large repository or history
  • Copying a proprietary codebase into public Git

Required inputs

  • Target directory (confirm before any write)
  • Intended project type, if the user already knows it
  • Whether this session may commit, add a remote, or push (default: no)
  • Chosen license, if the user already has one

Low-resource policy

Read the first of these that exists, then follow it:

  • ${CLAUDE_PROJECT_DIR}/.claude/shared/LOW_RESOURCE.md
  • $HOME/.claude/shared/LOW_RESOURCE.md

If neither exists, apply this fallback: run one expensive command at a time, prefer the narrowest validation, disable watch mode, reuse existing environments, and run full validation only at a milestone boundary. Do not scan the whole filesystem to locate the policy.

Context-efficiency policy

Read the first of these that exists, then follow it:

  • ${CLAUDE_PROJECT_DIR}/.claude/shared/CONTEXT_EFFICIENCY.md
  • $HOME/.claude/shared/CONTEXT_EFFICIENCY.md

If neither exists, apply this fallback: select files before reading; use targeted searches and bounded ranges; do not preload references; do not reread unchanged files; finish one atomic increment and stop; create a compact handoff before context is exhausted.

Facts that must not be assumed

  • Whether Git is already initialized
  • The default branch name
  • That the directory is empty
  • The stack, package manager, or test runner
  • That files with common names are safe to overwrite
  • That a remote exists or is empty

Preflight

  1. Confirm the absolute target directory with the user
  2. Run git rev-parse --is-inside-work-tree to detect an existing repo
  3. List top-level entries without reading large binaries
  4. If the directory already contains a substantial project, stop and hand off

to oss-plan or public-code-review

Workflow

  1. Confirm the target directory
  2. Detect Git state and existing files -> see

[references/project-detection.md](references/project-detection.md)

  1. Classify the project: Python, JavaScript or TypeScript, Rust, mixed stack,

documentation-only, or a Claude Code skills repository

  1. Identify existing manifests and conventions; do not replace them
  2. Define the smallest publishable structure for that stack ->

[references/stack-bootstrap.md](references/stack-bootstrap.md)

  1. Scan for public-release risks: secrets, personal information, proprietary

names, private URLs, accidental binaries, generated output

  1. Select or confirm license, README, .gitignore, .env.example (only when

secrets are required), and a minimal validation path

  1. Create only missing files; never overwrite an existing README, LICENSE, or

.gitignore

  1. Preserve executable permissions on existing scripts
  2. Prepare the staged diff and verify the exact file list ->

[references/git-first-publish.md](references/git-first-publish.md)

  1. Commit only when explicitly requested
  2. Add a remote only when explicitly requested
  3. Push only when explicitly requested
  4. Produce the report ->

[templates/bootstrap-report.md](templates/bootstrap-report.md)

Decision branches

  • Existing README, LICENSE, or .gitignore present -> keep it, record it as

preserved, and note any gap instead of rewriting

  • Secret or private data found -> stop, report the exact path and line, do not

stage it, do not commit

  • Binary or generated output found -> propose a .gitignore entry, do not add

it to the commit

  • No test or runnable path exists -> add the minimal validation described for

the stack; do not scaffold a framework

  • Git already initialized with commits -> do not re-initialize; treat as an

existing project

Validation escalation

Run the narrowest available check for the detected stack, in this order, only as far as needed to prove the repo is runnable:

syntax or parse check on created files
single smoke test or example run
stack lint or typecheck on changed files only

Do not run a full build or full suite during bootstrap.

Stop conditions

  • Target directory is ambiguous or unconfirmed
  • A secret or private datum would be committed
  • An existing README, LICENSE, or .gitignore would be overwritten
  • Git would be initialized in the wrong directory
  • The user has not authorized commit, remote, or push for those steps

Human review boundaries

  • Final license choice
  • Whether the project name and contents are cleared for public release
  • Remote URL and account ownership
  • Any file the risk scan flagged as uncertain

Final report

Fill in [templates/bootstrap-report.md](templates/bootstrap-report.md) with detected type, created files, preserved files, skipped files, validation performed, unresolved public-release risks, Git operations performed, Git operations not performed, and the next manual action.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.