Skip to content

hashgraph-online/awesome-codex-plugins / plugins/boshu2/agentops/skills/review

review

Verified

Give advisory feedback on a plan, design or code change. Use when: asked for an opinion or a look-over, even informally. Not for acceptance; use Validate.

FileKindSize
SKILL.mdskill md8.6 KB
references/advice-or-acceptance.mdmarkdown2.7 KB
references/document-findings.mdmarkdown1.7 KB

references/advice-or-acceptance.md

View on GitHub
# Advice, claim audit or acceptance

Review, Research's claimed-state audit and Validate share this routing rule.
Each keeps its own non-obvious rule inline; this page holds the shared detail,
so failing to read it never blocks advice, an audit or a judgment.

## Route by the intended outcome

Choose from what the caller wants to learn, not from the word "review" or
"check".

| The caller wants | Route | It returns |
|---|---|---|
| Suggestions, tradeoffs, concerns or a second look | [Review](../SKILL.md) | advisory findings with checked scope and gaps |
| To know whether a stated claim (done, shipped, fixed, goals met) matches the evidence | [Research claimed-state audit](../../research/SKILL.md#claimed-state-audit) | a per-claim disposition ledger, no verdict |
| An acceptance verdict, or independent proof that original acceptance is met | [Validate](../../validate/SKILL.md) | `PASS`, `FAIL` or `NOT_PROVEN` from a fresh context |

Explicitly selecting Validate, asking to establish that original acceptance is
met, or asking for independent proof of completion selects acceptance, even
when phrased as "review this".

## Ambiguous requests

Generic checking or readiness language ("can you check this?", "is it ready?")
selects none of the three by itself. A request for a look, once-over, opinion
or second look already names advice, even when the caller will act on the
result ("quick look before I close the ticket?"): return the findings and say
they are not acceptance. Supplied acceptance criteria say what to
inspect, not which kind of judgment the caller wants. When the request and
settled context leave the purpose open, ask one question: advisory findings or
an acceptance judgment? Wait for the answer. Do not return findings, a verdict
or a readiness conclusion while intent is unresolved. Missing intent is not a
`NOT_PROVEN` verdict.

A request that names a completion claim and asks whether it holds is not
ambiguous: it is a claim audit, so proceed without asking.

## Handing off to acceptance

When acceptance is wanted, stop the advisory route and give a genuinely fresh
Validate context the original acceptance, the exact subject, the complete
changed scope and pointers to the relevant evidence. Preserve required review
legs; Validate owns identity, freshness and verdict rules. A new role in the
current conversation is not a fresh context. If no fresh context or needed tool
is available, say what is missing and which handoff is needed; never claim that
validation occurred.

Advice, agreement, a claim audit or a no-finding result is never acceptance,
even when the caller asks for it to stand in for independent judgment. Clear
native work needs no mandatory skill or skill chain.
brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).Stars: 322kUpdated Claude CodeCodexVerifiedInstall
brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).Stars: 198kUpdated Claude CodeCodexVerifiedInstall