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 martineserios/thebrana --skill two-axis-reviewgit clone --depth 1 https://github.com/martineserios/thebranaWrote 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/martineserios/thebrana/two-axis-review)<a href="https://agentmods.dev/skills/martineserios/thebrana/two-axis-review"><img src="https://agentmods.dev/badge/skills/martineserios/thebrana/two-axis-review/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/martineserios/thebrana/two-axis-review"><img src="https://agentmods.dev/badge/skills/martineserios/thebrana/two-axis-review.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 | $0.00015 | $0.01764 |
| Opus 5 | $0.00008 | $0.00882 |
| Sonnet 5 | $0.00003 | $0.00353 |
| Haiku 4.5 | $0.00002 | $0.00176 |
Grade A, and why
two-axis-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 7d 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 — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Two-Axis Review (adapter)
ADR-084 vendor+wrap pilot expansion (t-2834's pilot evaluated EXPAND, §7a; t-2835 unparked under that same evaluation). Thin adapter, not a copy — the actual discipline lives in the vendored organ, read fresh every invocation (no state to lose across compaction — steps 1-3 below are the whole procedure):
Named two-axis-review, not code-review — deliberately distinct from the vendored
skill's own upstream name, because a pre-existing, unrelated built-in code-review skill
already occupies that name in this environment (correctness-bug + cleanup review, --fix/
--comment/--post flags). Vendoring the upstream organ verbatim at
.agents/skills/code-review/ (its name: frontmatter says code-review, per ADR-084 §2's
"lands verbatim") is unavoidable and correct — the collision is contained by giving this
adapter, the user/model-facing entry point, its own name, exactly as diagnose-hard-bug
already does for diagnosing-bugs (t-2834). Invoke the vendored organ directly only via this
adapter, never by typing /code-review and hoping for the Pocock skill.
-
Read the vendored skill:
Skill(skill: "code-review")—.claude/skills/code-review/SKILL.md, verbatim upstream, pinnedv1.2.3, tracked inskills-lock.json. Follow its 5-step process as written (pin fixed point → identify spec source → identify standards sources → spawn both sub-agents in parallel → aggregate under separate, never-merged## Standards/## Specheadings). -
While following it, remap the two inputs the upstream process expects the caller to supply:
Spec source (upstream step 2) → NOT
docs/agents/issue-tracker.md. That file is the tracker verb map (ADR-086 §8: "create an issue" →brana backlog add, etc.) — it answers "which CLI verb does X", not "what did this diff ask for". The actual spec brief for the Spec sub-agent comes from the originating task's own fields:- Resolve the task id from the branch name (
{epic}/{type}/t-NNN-{slug}, per this project's branch convention) or from whatever the user names explicitly. - Fetch the task (
backlog_get <t-NNN>) and runsystem/scripts/two-axis-spec-lookup.shagainst its JSON:acceptance_criteriawins if non-empty; otherwiseAC:-prefixed lines incontext(task-convention.md's "AC: prefix" convention); otherwise the script prints the literalno spec availableand exits 2. - Missing-spec case (AC4): exit 2 from the lookup script means the Spec sub-agent is
skipped entirely — the aggregated report says
## Spec\nno spec availableunder that heading, exactly as upstream step 2 already specifies for its own "user says there isn't one" branch. Never fabricate AC to fill the gap, and never fail the whole review because one axis has nothing to check — the Standards axis still runs and reports normally.
Standards source (upstream step 3) → brana's own documented conventions, not a
CODING_STANDARDS.md/CONTRIBUTING.mdfile (neither exists here):system/rules/*.md— the behavioural rules that bind all work regardless of language (sdd-tdd.md,no-patches-root-cause.md,universal-quality.md,git-discipline.md,parallel-bash.md, and any other rule file touching the diff's area).- Whichever language/domain skill applies to the diff (e.g.
brana:rust-skillsfor Rust,brana:bash-defensive-patternsfor shell) — pass its content inline to the Standards sub-agent the same way upstream expects aCONTRIBUTING.mdto be pasted in. - Plus the upstream skill's own Fowler smell baseline (its step 3, verbatim,
already ships in
.claude/skills/code-review/SKILL.md) — pasted in full alongside the brana-rules text, per upstream's own instruction that "the sub-agent has no other access to it." The baseline is a floor, not a ceiling: brana rules can suppress a baseline smell where they explicitly endorse the pattern it would flag (upstream's own "the repo overrides" clause, step 3).
- Resolve the task id from the branch name (
-
Cross-skill references → redirect-check.md, the committed list; re-verify it against the vendored
SKILL.mdon every upstream bump (ADR-084 §1 pump). The only slash-ref upstream carries (/setup-matt-pocock-skills) never fires — see the redirect list for why.
What ships with it
1 file 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.
- 7d ago First seen · 107 lines · 15 tokens per session scan A 232e80886990
two-axis-review is a skill published in the GitHub repository martineserios/thebrana (3 stars, last pushed 2d ago), licensed MIT. It adds 15 tokens to every session and 1,764 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 2026-09-03.
Other skills, from other repositories
triage-report
Drain the .review/ report queue. Resolves un-triaged scan/audit reports (producer-agnostic — check-docs-consistency and any other scanner), dispatches the triage-report subagent (Sonnet) one report at a time for per-finding dispositions (promote / patch / dup / needs-investigation / dismiss), then the gateway absorbs…
diffscope
A skill that turns recent code changes into a plain-language learning briefing. It shows what changed, how behavior differs, and explains relevant Python syntax.
crap-analyzer
Use to produce a risk-based refactor + test plan for recently-changed code on a diff/branch/PR by computing CRAP (complexity × untested) on changed methods. Multi-language — TypeScript, JavaScript, Python, Java, Kotlin, Go, Ruby, C#, Rust, PHP — auto-discovers how the repo generates coverage. Triggers …
github-commenting
How to post clean, rich, deduplicated GitHub PR review comments — suggestion blocks, multi-line anchors, markers, formatting rules. Load before posting or fixing any PR comment.
procoder
Work like a senior developer in a repository governed by procoder: run the commit gate before calling anything done, format and lint through the binary, and drive the spec, plan, todo, backlog, and sprint chain in .procoder/. Use this skill when the repository contains a .procoder/ directory or an AGENTS.md naming…
code-review
The depth half of a review - the dimensions a diff is read against (correctness, boundaries, concurrency, failure paths, secrets, data access, structure, test quality) and the rule that a finding is refuted before it is reported. The verdict stays with the reviewer agent. Use when reviewing a diff or a pull request…