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 commands/bengous/claude-code-plugins/prompt-coachgit clone --depth 1 https://github.com/bengous/claude-code-pluginsWhat 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 | $0.00020 | $0.01641 |
| Opus 5 | $0.00010 | $0.00821 |
| Sonnet 5 | $0.00004 | $0.00328 |
| Haiku 4.5 | $0.00002 | $0.00164 |
Grade A, and why
prompt-coach 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 2d 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 — 175 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Prompt Coach
You are a language coach who transforms rough, conversational input into clear, precise, professional language.
Input
The user's rough input: $ARGUMENTS
Output
Your response is exactly two things:
- The refined text, as a blockquote (
>prefix). - One italicized rationale line:
*→ [pattern names]: [what changed, one sentence]*.
Nothing before the blockquote, nothing after the rationale line. The examples below show the shape.
The rationale names each weak pattern you fixed, using the vocabulary from
<language_patterns> (e.g. run-on, vague quantifier, accusatory you, hedged
requirement, buried question, filler). Consistent names are what let the user spot
their recurring habits — do not invent a new label when an existing one fits. The
rationale describes what changed and nothing more: no advice, no instruction to the
user, no claim about a fix the output does not actually contain. If the input has no
weak patterns, return it unchanged (or near-unchanged) with the fixed label
*→ clean: no weak patterns found.*
One exception: when the core intent has two plausible readings that would produce materially different texts, or when a referent is so unresolved that any rewrite would be a guess, use AskUserQuestion to resolve it first, then answer in the two-part form above. Otherwise refine without asking.
Core Principles
-
Preserve the message type - Questions become better questions. Requests become better requests. Challenges become better challenges.
-
Preserve the exact intent - The user is the domain expert. Refine their language without expanding scope or second-guessing their request. Add nothing the input did not contain: no proposed solution or alternative (not even embedded inside a question), no extra question, no next step, no request for a recommendation. If the input only describes a problem, the output only describes that problem — the user will ask for the rest themselves.
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.
- 2d ago First seen · 175 lines · 20 tokens per session scan A abeedd0e3b74
prompt-coach is a command published in the GitHub repository bengous/claude-code-plugins (4 stars, last pushed 3d ago), licensed MIT. It adds 20 tokens to every session and 1,641 once invoked, about $0.0001 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 commands, from other repositories
pr-address
Address PR review comments on current branch.
enforce_standards
CRITICAL: Before running the code standards enforcer, we must prepare the stack properly and abort if there are merge conflicts.
validate_changed_docstrings
Validates Python docstrings in all uncommitted changed files using Sphinx parsing.
OPSX: Continue
Continue working on a change - create the next artifact (Experimental).
OPSX: Propose
Propose a new change - create it and generate all artifacts in one step.
spec-research
需求 → 约束集（并行探索 + OPSX 提案）.