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/shiyas1331/devkit/pr-reviewgit clone --depth 1 https://github.com/shiyas1331/devkitWhat 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 | $0.00035 | $0.01032 |
| Opus 5 | $0.00017 | $0.00516 |
| Sonnet 5 | $0.00007 | $0.00206 |
| Haiku 4.5 | $0.00003 | $0.00103 |
Grade A, and why
pr-review 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 — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PR Review Brief — router
This is a thin dispatcher. The picker fires when there's no flag; otherwise the command parses input and delegates to a sub-command. The full pipeline lives in commands/pr-review/default.md — the canonical body that other sub-commands reference.
Response format: one short sentence on what was done, the next concrete action, terse.
Input
$ARGUMENTS
Routing rules — apply in order
1. Picker fires when no flag is present
Trigger:
$ARGUMENTSis empty, OR$ARGUMENTScontains only a PR identifier (URL / number / branch) with no--*flag
Use AskUserQuestion. Chained questions: action then depth.
Question 1 — action:
question: "What do you want to do with the review?"
header: "Action"
multiSelect: false
options:
- label: "Read the brief"
description: "Print TL;DR + triage + decisions + risks in the terminal."
- label: "Save to disk"
description: "Write the brief to specs/reviews/PR-<num>-<slug>.md."
- label: "Post as a PR comment"
description: "Drop the whole brief as one comment on the GitHub PR. Confirms before posting."
- label: "Post as a full review (inline comments)"
description: "Native code-review style — comments at file:line with a summary body."
If $ARGUMENTS has no PR yet, prompt: "PR? (URL, number, or branch — e.g. 409 or feat/CAT-260-foo)".
Question 2 (only if "Read" or "Save") — depth:
question: "How much depth?"
header: "Depth"
multiSelect: false
options:
- label: "Full brief (default)"
description: "Includes TL;DR, triage table, key decisions, conventions, risks."
- label: "Quick TL;DR only"
description: "Just the headline + triage table."
Map answer → sub-command and re-invoke:
| Action | Depth | Re-invoke |
|---|---|---|
| Read | Full | /devkit:pr-review <PR> (default sub-command) |
| Read | Quick | /devkit:pr-review <PR> --depth=quick |
| Save | Full | /devkit:pr-review <PR> --save |
| Save | Quick | /devkit:pr-review <PR> --depth=quick --save |
| Post comment | (skip Q2) | /devkit:pr-review <PR> --post |
| Post review | (skip Q2) | /devkit:pr-review <PR> --post-review |
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 · 102 lines · 35 tokens per session scan A f61a1ee4c598
pr-review is a command published in the GitHub repository shiyas1331/devkit (4 stars, last pushed 1mo ago), licensed MIT. It adds 35 tokens to every session and 1,032 once invoked, about $0.0002 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
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.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
constitution
Create or update the project constitution from interactive or provided principle inputs.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.