ai-startup-skills
创业公司做 AI 产品的 7 步 Agent Skill 合集:验证 → BP → 政策 → 路演 → 获客 → 定价 → 注册。
Translation pending
Working builds made by asking AI: agents, guardrails, desktop tools, scheduled jobs and Salesforce delivery.
OtherPythonGitHub @HenryLopOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
A credit guard for paid generation APIs: runs start as requests with model and credit cost, a typed approval naming the count arms a single run, the guard relocks after, and a ledger records cost per run.
Taken from the repository README
Loading the full README.
Read the full README ↗A credit guard for paid generation APIs: runs start as requests with model and credit cost, a typed approval naming the count arms a single run, the guard relocks after, and a ledger records cost per run.
python demo.py
Runs the whole gate end to end against the real hooks and CLI, in a temp state dir. Tests: python -m unittest discover -s tests.
Python 3.8+. No dependencies.
python -m creditguard request "intro clip" --model video-x --settings "720p 5s" --cost 12.approve hook (UserPromptSubmit) reads the prompt; typing go 12 approves the pending request that costs exactly 12 credits. A wrong count approves nothing, and a negated phrase ("do not go 12") is ignored.python -m creditguard arm <id> --quoted <cost> arms exactly one run, and refuses if the quoted price differs from what was approved.guard hook (PreToolUse) denies any generate command unless an arm is live, and spends the arm on first use, so a retry is locked again. Arms expire after 5 minutes.python -m creditguard ledger prints the total. Optional CREDIT_BUDGET caps cumulative spend.Hook wiring for Claude Code (settings.json):
{"hooks": {
"UserPromptSubmit": [{"hooks": [{"type": "command", "command": "python creditguard/hooks.py approve"}]}],
"PreToolUse": [{"matcher": "Bash", "hooks": [{"type": "command", "command": "python creditguard/hooks.py guard"}]}]
}}
Set CREDIT_GENERATE_PATTERN to a regex matching your tool's paid command.
Measured from the author's own use of the guard around a paid video and image service, totals only, no prompts or project names. Requests were filed 20 to 23 Oct 2026.
| Measure | Count |
|---|---|
| Requests filed | 32 |
| Approved by a typed approval | 20 |
| Never approved | 12 |
| Approved requests that ran | 13 |
| Approved requests that never ran | 7 |
MIT. See LICENSE.
Built with Claude Code. More: https://made-by-asking.pages.dev/showcase/studio-credit-guard/