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 hf745316389/ho-codeflow --skill ho-reviewgit clone --depth 1 https://github.com/hf745316389/ho-codeflowWrote 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/hf745316389/ho-codeflow/ho-review)<a href="https://agentmods.dev/skills/hf745316389/ho-codeflow/ho-review"><img src="https://agentmods.dev/badge/skills/hf745316389/ho-codeflow/ho-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/hf745316389/ho-codeflow/ho-review"><img src="https://agentmods.dev/badge/skills/hf745316389/ho-codeflow/ho-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 | $0.00039 | $0.01396 |
| Opus 5 | $0.00019 | $0.00698 |
| Sonnet 5 | $0.00008 | $0.00279 |
| Haiku 4.5 | $0.00004 | $0.00140 |
Grade A, and why
ho-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 10d 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 — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review
Decide whether the change meets its acceptance criteria, from the artifacts themselves. You are not fixing anything in this phase.
What you read
change.yaml, 01-design.md, 02-implementation.md, the actual files the
change touched, and the output of the verification commands run now, by you.
The implementation report tells you where to look and what the implementer believes. It is a map, not evidence. Every result you record comes from a file you opened or a command you ran in this session.
More than one round
02-implementation.md carries the current round first and earlier rounds below
under ## Round N (superseded). Read the current round for what to check, and
the earlier ones for what was already accepted or left unverified.
If an earlier round is missing — replaced rather than kept — say so. You are reviewing round N without the record of round N-1, and that limits what you can conclude about deviations carried forward.
If 01-design.md has an Amendments section, check that each amendment
matches what the user actually said and does not quietly widen scope.
One result per criterion
Every acceptance criterion in the design gets exactly one of:
| Result | Means |
|---|---|
pass |
You checked it and it holds |
fail |
You checked it and it does not hold |
unverified |
You could not check it, and you say why |
There is no fourth value. Not partial, not mostly, not n/a. A criterion
the implementer did not attempt is fail, not unverified — unverified is
about your ability to check, not about their decision to skip.
If the design offered the user alternatives and their answer picked one, the
criteria belonging to the options they did not pick are not criteria of this
change. They should have been removed from the design when the amendment
landed; if they are still there, say so under Artifact integrity and score
only the ones in force.
unverified is a legitimate outcome when the check needs a fixture, production
data, or an external system you do not have. It is not an automatic failure.
Whether the change can still be complete with an unverified item is settled
by the design's own definition of done, not by you.
What ships with it
2 files 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.
- 10d ago First seen · 140 lines · 39 tokens per session scan A b095ac89bac1
ho-review is a skill published in the GitHub repository hf745316389/ho-codeflow (4 stars, last pushed 27d ago), licensed Apache-2.0. It adds 39 tokens to every session and 1,396 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 skills, from other repositories
issue-creation
Trigger: issue creation, bug reports, feature requests, or issue approval. Create and triage GitHub issues from repository evidence.
sdd-tasks
Break an SDD change into implementation tasks. Trigger: orchestrator launches task planning for a change.
work-unit-commits
Plan commits as reviewable work units. Trigger: implementation, commit splitting, chained PRs, or keeping tests and docs with code.
rework-rate
Measure and interpret PR rework rate — the emerging 5th DORA metric.
systemic-issue-triage
Trigger: new issue, bug report, triage, backlog, issue flood, community report, root cause, dead-end, blocked user. Attack issues by root class, never one-by-one; fixes must shrink the system, not grow it.
sdd-research
Trigger: SDD research, external evidence, source-backed research. Produce auditable evidence for a selected research lane.