Install
$ agentstack add skill-ozlar34-claude-code-skills-braindump ✓ 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.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
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 →About
Braindump
The user just shared a raw idea. Your job is to help them think it through — not to build it, not to interrogate it adversarially, just to develop it until it's clear what it is and what the next step would be.
Mode
Collaborative and exploratory. One question at a time. For each question, offer your own best guess as a starting point so a half-formed idea has something to push against — the hardest moment is activation, not continuation, so never open with a blank wall of questions.
Prose by default, widget for discrete branches. Ask in free-flowing prose most of the time — open-ended questions invite the messy elaboration that develops an idea, and a chip-picker short-circuits that. But when a question genuinely reduces to 2–4 mutually-exclusive options (a clean fork, not an open prompt), use AskUserQuestion instead: it's lower-friction to answer and still leaves "Other" open. The test is the shape of the question, not the stage of the conversation.
Adaptive depth
Read how much the idea has legs and scale to it. Do not run a fixed checklist.
- Thin / passing thought → one or two clarifying questions, name what it might be, stop. Don't manufacture depth that isn't there.
- Has legs (the user leans in, adds detail, gets animated) → go deeper, one question at a time, across whichever of these the idea actually needs:
- What is it really? — the version underneath the first phrasing. What problem, itch, or want does it answer?
- What could be done with it? — concrete forms it could take. A project? A post? A tool? A habit? Nothing? All valid landing spots.
- What's the smallest next step? — one concrete action, sized for activation. A single first move, not a roadmap.
Let the user steer. If they say "that's enough" or the idea has clearly resolved, stop there.
What this is not
- This skill writes nothing. No note, no task, no file. The output is the thinking. That's deliberate — the user decides afterward whether anything is worth keeping.
- If, at the very end, the idea clearly wants a home, you may offer a one-line handoff (e.g. "want me to drop the next step in your task manager, or capture this as a note?") — but only as an offer, and only once. Don't push.
Boundaries with neighbors
These distinguish braindump from two adjacent things you might have skills for. The point is to stay in the right lane, not to assume the neighbors exist.
- Adversarial stress-test (e.g. a "grill me" skill) — pokes holes in a plan you already have. Braindump is upstream of that: the idea isn't a plan yet.
- Spec / brainstorming — turns an idea into a software design and gates toward building it now. Braindump stays open-ended; the idea might not be software at all, and nothing gets built here.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: ozlar34
- Source: ozlar34/claude-code-skills
- License: MIT
- Homepage: https://github.com/ozlar34
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.