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.
git clone --depth 1 https://github.com/tslateman/duetWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/tslateman/duet/ponder)<a href="https://agentmods.dev/commands/tslateman/duet/ponder"><img src="https://agentmods.dev/badge/commands/tslateman/duet/ponder/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/tslateman/duet/ponder"><img src="https://agentmods.dev/badge/commands/tslateman/duet/ponder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00006 | $0.00373 |
| Opus 5.5 | $0.00002 | $0.00149 |
| Sonnet 5.5 | $0.00001 | $0.00075 |
| Haiku 4.5 | $0.00001 | $0.00037 |
Grade A, and why
ponder 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 1mo ago.
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.
What it actually says
Pick ${ARGUMENTS:-1} question(s) at random from this list and present them:
- What would a 'Navigation Dead Reckoning' system look like for career decisions?
- How do you architect for forgetting - designing systems that gracefully degrade memories they no longer need?
- What's the engineering equivalent of muscle memory, and how do you deliberately train it?
- What would a 'compiler for human intentions' look like?
- What's the half-life of a decision, and how do you build with decay in mind?
- What would 'emotional CI/CD' look like?
- How do you measure growth through adversity without incentivizing unnecessary suffering?
- What patterns in your codebase would embarrass you if they became visible?
- How do you design for the engineer you'll be in 5 years, not the one you are today?
- What's the 'tidal' pattern in your work - the thing that advances and retreats but never stops?
- What would 'earthquake-ready' architecture look like for sudden market shifts?
- How do you instrument for questions you don't yet know how to ask?
- What's the simplest complex thing you've built, and the most complex simple thing?
- How do you paint with constraints - making limitation a creative force?
- What would controlled burns look like in software - deliberate destruction for ecosystem health?
- If 'people arrive at beliefs based on what they find attractive' (Pascal), how do you make good architecture attractive?
- What technical truths do you think you possess directly, and which might be mere familiarity?
- How do you design systems that please as much as they convince?
- What beliefs about your codebase exist not from proof but from attraction?
- When you consider the infinite space of possible solutions, which silences alarm you?
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.
- 1mo ago First seen · 28 lines · 6 tokens per session scan A 332f65863ef6
ponder is a command published in the GitHub repository tslateman/duet (1 stars, last pushed 1mo ago), licensed MIT. It adds 6 tokens to every session and 373 once invoked, about $0.0000 per session on Opus 5.5. 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 2026-08-31.
Other commands, from other repositories
quick
Quick execution mode for ad-hoc tasks without full workflow overhead.
audit-release
Diagnose release-readiness against the Starter Series quality bar — matched starter, version-vs-last-tag drift, CHANGELOG drift vs merged PRs, and publish-workflow kind. Read-only.
recap
Time-windowed "where was I?" — git commits, session notes, and files touched in the last N hours or days. Use after stepping away for a few hours or days. Not for cold starts after a long absence (use /tools:brief) or live development checks (use /tools:status).
status
Warm mid-session check — git detail, last session note, and in-progress backlog. Use when you still have mental context loaded and want a quick pulse. Not for cold starts (use /tools:brief) or planning (use /tools:overview).
wrap
You are closing a Claude Code session. Execute all steps below in order.
workday-complete
End-of-day wrap — validate, consolidate, review, changelog.