Install
$ agentstack add skill-nobrainer-tech-nobrainer-claude-skills-nobrainer-autopilot ✓ 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 No
- ✓ 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
NoBrainer Autopilot
Autonomous CI/CD workflow for any project. Converts TODOs to issues, spawns expert teams, implements, creates PRs with Copilot review gate, merges — all hands-free.
Trigger: "autopilot", "go go", "dzialaj sam", "lec sam", "cisnij", "autopilot mode", "autonomous mode", "work autonomously"
Prerequisites Check
Before starting, verify:
- nobrainer-team-builder installed:
``bash ls ~/.claude/skills/nobrainer-team-builder/skill.md ` If missing: git clone https://github.com/nobrainer-tech/nobrainer-claude-skills.git /tmp/nbs && cp -r /tmp/nbs/nobrainer-team-builder ~/.claude/skills/`
- GitHub Copilot review configured on the repo:
``bash # Check if Copilot reviews exist on recent PRs gh api repos/{owner}/{repo}/pulls?state=closed\&per_page=5 --jq '.[].number' | while read pr; do has=$(gh api repos/{owner}/{repo}/pulls/$pr/reviews --jq '[.[] | select(.user.login == "copilot-pull-request-reviewer[bot]")] | length') if [ "$has" -gt 0 ]; then echo "Copilot active on PR #$pr"; break; fi done `` If not configured: warn user to enable Copilot code review in repo settings → Code review → Copilot.
- CLAUDE.md exists in project root with coding standards.
Phase 1: Discovery & Planning
Step 1: Codebase Review
Spawn 3 agents in parallel:
- code-explorer: Map architecture, key files, dependencies
- code-reviewer: Find bugs, code quality issues, missing tests
- test-engineer: Assess test coverage gaps
Step 2: Collect TODOs
Gather work items from ALL sources:
# Existing GitHub issues
gh issue list --state open --json number,title,labels
# TODO/FIXME/HACK in code
grep -rn "TODO\|FIXME\|HACK\|XXX" --include="*.py" --include="*.ts" --include="*.js" .
# tasks/todo.md if exists
cat tasks/todo.md 2>/dev/null
Step 3: Present Plan to User
Show a prioritized list:
## Autopilot Plan
| # | Priority | Source | Description |
|---|----------|--------|-------------|
| 1 | HIGH | Issue #42 | Fix auth bypass vulnerability |
| 2 | HIGH | Code review | 3 files over 300-line limit |
| 3 | MED | TODO | Add input validation to API |
| 4 | LOW | Issue #38 | Refactor logging module |
Proceed on autopilot? (adjust priorities or say "go")
Step 4: Convert to Issues
After user approves, ensure each item has a GitHub issue:
# Create issues for TODOs not yet tracked
gh issue create --title "..." --body "..." --label "autopilot"
Phase 2: Autonomous Execution Loop
For each issue (highest priority first):
Step 1: Spawn Expert Team (min 5 agents)
Use nobrainer-team-builder skill with minimum 5 agents. Team composition based on task:
| Task Type | Agents | |-----------|--------| | Bug fix | python-pro, debugger, test-engineer, code-reviewer, security-auditor | | Feature | python-pro, architect-reviewer, test-engineer, code-reviewer, debugger | | Refactor | refactoring-specialist, code-reviewer, test-engineer, python-pro, architect-reviewer | | Security | security-auditor, python-pro, test-engineer, code-reviewer, debugger |
Step 2: Implement
- Create feature branch from main
- Implement the fix/feature
- Run tests:
python -m pytest tests/ -x - Verify coding standards:
- DRY — no duplicated logic
- KISS — simplest solution that works
- SOLID — single responsibility, open/closed
- YAGNI — no speculative features
- Max 300 lines per file (up to 350 acceptable)
- Max 80 lines per function
Step 3: Create PR
git push -u origin {branch}
gh pr create --title "..." --body "Fixes #{N}\n\n## Summary\n...\n\n## Test plan\n..."
Step 4: Copilot Review Gate
CRITICAL — the most important step. NEVER skip.
# Poll until Copilot review EXISTS (3-5 min typical)
for i in $(seq 1 10); do
reviews=$(gh api repos/{owner}/{repo}/pulls/{N}/reviews \
--jq '[.[] | select(.user.login == "copilot-pull-request-reviewer[bot]")] | length')
if [ "$reviews" -gt 0 ]; then echo "Copilot reviewed"; break; fi
echo "Waiting for Copilot... ($i/10)"
sleep 60
done
After review exists:
- Read inline comments:
gh api repos/{owner}/{repo}/pulls/{N}/comments - Reply to EACH comment (1-2 sentences)
- Resolve EACH thread via GraphQL
- If fixes needed → commit, push, wait for re-review (repeat from poll)
- Verify 0 unresolved threads:
``bash gh api graphql -f query='{ repository(owner:"{owner}", name:"{repo}") { pullRequest(number:{N}) { reviewThreads(first:50) { nodes { isResolved } } } } }' \ --jq '[.data.repository.pullRequest.reviewThreads.nodes[] | select(.isResolved==false)] | length' ` Must return 0`.
Step 5: Merge
gh pr merge {N} --squash --delete-branch
Step 6: Deploy (if configured)
If project has deploy instructions in CLAUDE.md → deploy after merge.
Step 7: Next Issue
Go back to Step 1 with next priority issue.
Quality Gates (enforced on every PR)
| Gate | Check | |------|-------| | Tests pass | pytest -x exit 0 | | No file > 350 lines | find . -name "*.py" -exec wc -l {} \; \| awk '$1>350' | | No function > 80 lines | Code review agent | | DRY/KISS/SOLID/YAGNI | Code review agent | | Copilot reviewed | Review exists + 0 unresolved threads | | CI passes | gh pr checks {N} all pass |
Rules
- Min 5 agents per task — diminishing returns below that for quality
- Always create issue first — never implement without tracked issue
- Copilot gate is NON-NEGOTIABLE — even for "trivial" changes
- Reply before resolve — never silently resolve threads
- 300-line limit — up to 350 acceptable, over 350 must split
- Deploy after merge — if deploy instructions exist in CLAUDE.md
- Ask user only when blocked — autopilot means autonomous
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: nobrainer-tech
- Source: nobrainer-tech/nobrainer-claude-skills
- License: MIT
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.