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 skills add afaraha8403/balakit --skill deep-deliberationgit clone --depth 1 https://github.com/afaraha8403/balakitWrote 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/skills/afaraha8403/balakit/deep-deliberation)<a href="https://agentmods.dev/skills/afaraha8403/balakit/deep-deliberation"><img src="https://agentmods.dev/badge/skills/afaraha8403/balakit/deep-deliberation/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/skills/afaraha8403/balakit/deep-deliberation"><img src="https://agentmods.dev/badge/skills/afaraha8403/balakit/deep-deliberation.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 | $0.00060 | $0.01898 |
| Opus 5 | $0.00030 | $0.00949 |
| Sonnet 5 | $0.00012 | $0.00380 |
| Haiku 4.5 | $0.00006 | $0.00190 |
Grade A, and why
deep-deliberation 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 8d 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.
How it starts
The opening of the file, as written. The whole thing — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Deep Deliberation
Leading words: option tree, evidence tournament, reversibility, adversarial review, premortem, uncertainty, checkpoint, ground truth.
Deep Deliberation frames a consequential decision, compares viable alternatives, tests them against evidence, and adjudicates unresolved risks. The pipeline produces a decision record, not implementation.
Operating contract
- Do not edit files, implement code, or create formal plan documents while the pipeline is active.
- Ground claims in repository evidence and current external sources when the decision depends on them.
- Keep at least two viable approaches through Stage 2 to resist early anchoring.
- Preserve material disagreements; never silently average them away.
- Treat isolated delegates as separate contexts, not independent authorities.
- Never invent missing delegate findings or unsupported evidence.
- Stop after every checkpoint. Continue only after the user explicitly responds.
- The user may revise, restart, or end the pipeline at any checkpoint.
- A meta question does not advance the state; answer it, repeat the state block, and remain at the current checkpoint.
Runtime state
Start every pipeline response with this block and update it mechanically:
DELIBERATION_STATE
stage: 1 | 2 | 3
checkpoint: none | 1 | 2 | 3
shortlist: unset | A,B
delegate_results: 0/N
next_action: one action only
If the state is missing after a context change, reconstruct it from the latest completed template and ask the user to confirm before advancing.
Pipeline
- Stage 1 — Frame and shortlist: define the decision, generate 3–5 approaches, and shortlist the strongest two.
- Checkpoint 1: user approves the framing and shortlist.
- Stage 2 — Evidence tournament: 3–5 focused reviewers compare both shortlisted approaches using a normalized evidence contract.
- Checkpoint 2: user reviews findings and unresolved disputes.
- Stage 3 — Adjudication: targeted evidence review and premortem produce the final recommendation.
- Checkpoint 3: user approves, revises, or rejects the recommendation.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 8d ago First seen · 238 lines · 60 tokens per session scan A abce6f4b01af
deep-deliberation is a skill published in the GitHub repository afaraha8403/balakit (1 stars, last pushed 11d ago), licensed MIT. It adds 60 tokens to every session and 1,898 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 2026-08-31.
Other skills, from other repositories
fable-thinking
Reasoning protocol distilled from Claude Fable 5.1. Makes any model reason like Fable — evidence-grounded claims, multi-hypothesis diagnosis, concrete simulation, adversarial self-review, calibrated outcome-first delivery. Its never-skipped Floor check catches simple-looking trick questions models answer confidently…
template-pipeline
Numbered pipeline stage orchestrators (stages 00–13) for the template research framework. Each stageNN.py is a thin orchestrator that wires the repo root onto sys.path and delegates to infrastructure/ modules.
template-runner
Pipeline execution runners for the template research framework. Contains executepipeline.py (single-project), executemultiproject.py (multi-project serial/parallel), and runmatrix.py (deterministic matrix).
expert-redteam-review
Use when a user asks to evaluate a complex, high-impact, ambiguous, cross-functional, or hard-to-reverse decision; mentions expert panel, red team, adversarial review, rebuttal, judge, arbitration, risk review, challenge assumptions, poke holes, or wants stronger decision quality than a normal review.
phase-enforcement
Evidence-based phase completion enforcement for /dev:feature. Use when orchestrating 8-phase feature development to ensure artifacts exist before phase completion, validation criteria are addressed, outer loops are enforced, and show-your-work requirements are met.
pm-verification-protocols
QA verification gate and evidence requirements.