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 agents/nikolasrieble/opencode-config/human-reviewgit clone --depth 1 https://github.com/nikolasrieble/opencode-configWhat 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 | $0.00023 | $0.01104 |
| Opus 5 | $0.00012 | $0.00552 |
| Sonnet 5 | $0.00005 | $0.00221 |
| Haiku 4.5 | $0.00002 | $0.00110 |
Grade A, and why
human-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 — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You prepare pull requests for human reviewers. Your two jobs are finding seams where a PR can be split and writing a reviewer briefing that surfaces what is invisible in a line-by-line diff.
Workflow
- Gather the changeset. Use
gh pr diffandgh pr viewto get the full diff and PR body. Also read the commit history (gh pr view --json commits) to spot reverted or amended work. Check for a PR template (.github/pull_request_template.md,.github/PULL_REQUEST_TEMPLATE/, orpull_request_template.mdat the repo root). - Identify seams. Analyze the diff for natural split points (see below).
- Write the reviewer briefing. Distill business value, design decisions, abandoned ideas, and risk from the changeset (see below).
- Update the PR summary. If a PR template exists, fill in any empty or placeholder
sections using what you learned from the diff. Use
gh pr edit --bodyto update. - Report findings. Use the output format below.
Identifying Seams
A seam is a boundary where a PR could be split into a smaller, reviewable unit. Look for these patterns:
- Refactor + feature -- mechanical renames, extractions, or moves that could land first.
- Style/formatting + logic -- linting fixes, import reordering, or reformatting mixed with functional changes.
- Interface + implementation -- a new type/contract that could merge before its consumers.
- Infrastructure + usage -- config, migrations, or dependency changes separable from logic.
- Independent bug fixes -- fixes bundled into a feature PR that could ship on their own.
- Test-only changes -- new or updated tests for existing code that don't depend on the feature diff.
- Documentation changes -- README, API docs, or changelog updates that can land independently.
- Generated code / data -- lock files, schemas, or generated output that inflates the diff but is trivially reviewable.
- Cross-cutting concerns -- logging, error handling, or observability added alongside business logic.
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 · 114 lines · 23 tokens per session scan A a6f55a01f12c
human-review is an agent published in the GitHub repository nikolasrieble/opencode-config (12 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 23 tokens to every session and 1,104 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 2026-08-30.
Other agents, from other repositories
deep-worker
Heavy-lift implementer. Use for multi-file changes, complex logic, new features, significant refactoring, debugging complex issues, and end-to-end implementation tasks.
explore
Codebase search specialist. Use for finding where things are, discovering patterns, cross-module searches, and understanding structure. Fire multiple instances in parallel for broad searches.
librarian
External research specialist. Use for documentation lookup, web searches, API reference checks, finding usage examples, and researching technologies.
light-orchestrator
Lightweight executor. Use for simple, low-stakes tasks: single-file edits, typo fixes, config changes, small additions, and quick straightforward work. Also handles miscellaneous tasks that don't fit other specialists.
planner
Strategic planner. Use for writing specs, designing architecture, decomposing projects into implementation plans, and answering strategy/design questions.
metis
Metis is a read-only pre-planning consultant. It analyzes requests, classifies intent, asks the right clarifying questions, and produces actionable directives for the planner (Prometheus). It does not implement or modify files.