Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/quantumbyteoss/quantumbyte/disqualificationnpx skills add QuantumByteOSS/quantumbyte --skill disqualificationgit clone --depth 1 https://github.com/QuantumByteOSS/quantumbyteWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00058 | $0.00544 |
| Opus 5 | $0.00029 | $0.00272 |
| Sonnet 5 | $0.00012 | $0.00109 |
| Haiku 4.5 | $0.00006 | $0.00054 |
Grade A, and why
disqualification scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 31 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Disqualification — the honest exit
When the fit is genuinely wrong, building something that won't land wastes the user's time and burns trust. The honest exit is the right move — and a last resort, never the first. Try a default sketch first; only disqualify when the mismatch is structural.
When it fires
- Realism failure — the ask is outside what QB does well (e.g. a consumer social network at megascale, a hardware product, something requiring a team and capital QB can't substitute for).
- Sustained low magnitude — no real business, no stakes, no agency, after the opening loop genuinely tried.
Note: this is about mismatch, not silence. A cold-start user who simply hasn't engaged is a soft re-engage, not a disqualification.
How to do it
- Be straight, early. Name the mismatch in plain terms, without hedging or fake humility. "I'll be straight: QB ships working operational tools for a specific business, fast. A at isn't something I'd do well, and I won't pretend otherwise."
- Find the buildable core, if there is one. Often a piece of the idea is in scope. Offer it. "If there's an operational tool inside this — managing , tracking — I can build that today."
- Point them somewhere real if there isn't. Suggest the honest alternative path (a different category of tool, a different approach).
- Preserve dignity. Don't chase. No guilt, no "are you sure?", no repeated pitch. Reset the frame to "this isn't the right fit, here's what is."
What this protects
A clean honest exit costs you this user but protects the brand and the trust signal — and frequently surfaces the in-scope sub-problem that becomes a good build. Pretending to qualify an unqualifiable ask produces a generic app, a cold reveal, and a churned, distrustful user.
Dependencies (fall back if absent)
Needs the qualification read to trigger. If qualification isn't running, this skill simply doesn't fire — the spine proceeds and <honesty> in the system prompt still lets the agent push back on an out-of-scope ask. No regression.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 31 lines · 58 tokens per session scan A 92e29bbfff13
disqualification is a skill published in the GitHub repository QuantumByteOSS/quantumbyte (323 stars, last pushed 19d ago), licensed Apache-2.0. It adds 58 tokens to every session and 544 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
gsd
Get Shit Done (GSD) - A comprehensive project management system for solo developers using Claude agents.
st-full-workflow
Use when the user asks to run the complete end-to-end Strikethroo workflow for a work order in one shot in this repository — triggers include full workflow, end-to-end, plan and execute, do everything, run the whole strikethroo workflow. Do not use when the user wants only one stage (create a plan, generate tasks, or…
st-execute-blueprint
Use when the user asks to run, execute, implement, or carry out a Strikethroo plan or its blueprint by plan ID in this repository — triggers include execute blueprint, run the plan, implement plan, build the plan. Do not use to create a plan, to only generate tasks, to run a single task, or for generic development…
st-generate-tasks
Use when the user asks to decompose, break down, or generate tasks for an existing Strikethroo plan ID in this repository — triggers include generate tasks, break down the plan, decompose plan, create the task blueprint. Do not use to create a new plan, to execute tasks, or for generic project planning outside…
st-code-review
Use when the blueprint execution gate asks for an independent second-harness review of a Strikethroo plan's cumulative diff in this repository — triggers include code review gate, review the plan diff, second-model review, CODEREVIEW hook, review the cumulative diff. Do not use to review a single task, to review code…
st-execute-task
Use when the user asks to run, execute, or implement one specific task ID within a Strikethroo plan in this repository — triggers include execute task, run task, implement task, do task N. Do not use to execute a whole plan or blueprint, to generate tasks, or for generic development outside Strikethroo.