Borrowing it
Nothing to install: this file belongs to future-architect/uzomuzo-oss. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/future-architect/uzomuzo-oss/main/.github/prompts/plan-debate.prompt.mdgit clone --depth 1 https://github.com/future-architect/uzomuzo-ossWrote 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/future-architect/uzomuzo-oss/plan-debate)<a href="https://agentmods.dev/commands/future-architect/uzomuzo-oss/plan-debate"><img src="https://agentmods.dev/badge/commands/future-architect/uzomuzo-oss/plan-debate/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/future-architect/uzomuzo-oss/plan-debate"><img src="https://agentmods.dev/badge/commands/future-architect/uzomuzo-oss/plan-debate.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 | $0.00000 | $0.09282 |
| Opus 5.5 | $0.00000 | $0.03713 |
| Sonnet 5.5 | $0.00000 | $0.01856 |
| Haiku 4.5 | $0.00000 | $0.00928 |
Grade D, and why
plan-debate scanned grade D with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
chmod 0700 "$DEBATE_DIR" || die_clean "ERROR: chmod debate dir failed" Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf "$DEBATE_DIR" || echo "WARN: failed to remove $DEBATE_DIR (holds a plan copy; rm it manually)" >&2 How it starts
The opening of the file, as written. The whole thing — 530 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/plan-debate — gpt-5.6-sol review → Claude↔gpt-5.6-sol debate → architect arbitration
This skill is the intermediate stage in Claude (Opus, drafts a plan in plan mode) → this skill → user check → ExitPlanMode + implement.
/plan-review is "gpt-5.6-sol critiques the plan once and stops". This skill is the heavyweight version, adding three things:
- Rebuttal debate — Claude rebuts / concedes each gpt-5.6-sol finding, and gpt-5.6-sol re-evaluates, for up to two rounds. This kills gpt-5.6-sol false positives based on misreading, and deepens the genuine findings.
- Neutral arbitration — after the debate, the
architectsubagent rules "adopt / revise / reconsider" using the repo's DDD rules and ADRs. This avoids the bias of the thread that wrote the plan grading its own work. - Cross-vendor perspective — the plan author is Claude (Opus, Anthropic); the critic is gpt-5.6-sol (OpenAI, via Copilot CLI). Two models from different vendors are less likely to share the same blind spot; we run their conflict as an arbitrated debate rather than a single reviewer.
Mode
/plan-debate(no args) — target the latest plan file in/home/node/.claude/plans//plan-debate <path>— target the given plan file (only paths under the plans dir; path traversal / symlink /-prefix / non-.mdrejected)- no plan file → return
No plan file foundand exit
When to use / not to use
When: before ExitPlanMode on an important / high-rework-cost plan (new package, layer-structure change, cross-layer refactor, a design decision that warrants an ADR). When one critique is not enough and you want a cross-vendor perspective plus a repo-rule-grounded ruling.
When NOT:
- a trivial plan (typo / comment / single flag) —
/plan-review's single sanity check is enough; debate is overkill - diff review (still at the plan stage; post-implementation diff is
/review-diff) - offline /
gh authnot completed - AI Credits budget is tight (this skill calls copilot 2-3 times, so 2-3x
/plan-review— see Cost) - the plan contains NDA / secret content (see Trust boundary)
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 · 530 lines · 0 tokens per session scan D 54207daa444c
plan-debate is a command published in the GitHub repository future-architect/uzomuzo-oss (42 stars, last pushed today), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 9,282 tokens. A static security scan graded it D with 2 findings (asks for root, recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-07.
Other commands, from other repositories
format
Format a SQL string or .sql file with GoSQLX.
security
Scan a SQL string, .sql file, or SQL-building source file for injection risks with GoSQLX.
analyze
Full GoSQLX analysis of a SQL string or .sql file (validity, metadata, security, lint, formatting).
lint
Lint a SQL string or .sql file with GoSQLX style rules (L001-L010).
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.