risk-review
Skill0 starsChanged 2 months agoMITView on GitHub
Review one B2B SaaS account for supported risk mechanisms, alternative explanations and consequential unknowns. Use when investigating an account-risk concern, challenging a health alert or deciding whether customer intervention or evidence verification is appropriate. Not for generic account summaries, portfolio ranking, health scores or churn probabilities.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- tyabu12/claude-kittheir own version, first seen Oct 20, 2026
- madeordinary/serel-memorytheir own version, first seen Oct 9, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- madeordinary/serel-memory★ 1same name
- tyabu12/claude-kit★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Read the job contract and evidence rules before reviewing the account.
- Distinctive: line 18
Use evidence for one identifiable account and a stated as-of date.
- Distinctive: line 49
Every consequential claim is traceable.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
risk-review
Decision: What does the evidence support, what remains unknown, and what should happen next?
Read the job contract and evidence rules before reviewing the account.
Inputs
Use evidence for one identifiable account and a stated as-of date. Accept notes, excerpts, documents or structured records; do not require a particular packet format. Establish the customer objective, workflow and relevant timing from the evidence, or mark them unknown.
Company policies, escalation routes and owner mappings are optional. They can refine recommendations but are not prerequisites for a qualitative finding. A CRM health label is a claim to investigate, not ground truth.
Workflow
- Establish account identity, cutoff and source coverage. Ask for missing identity or timing when it prevents a defensible assessment; otherwise state the limit and proceed only within the supported scope.
- Separate observations, customer statements, interpretations and gaps. Exclude wrong-account and future evidence; surface stale or conflicting evidence rather than silently resolving it.
- Connect each concern to a customer objective or critical workflow and explain the possible failure mechanism. If that connection is unknown, label it as a hypothesis—not an established business impact. Usage, ticket volume and sentiment alone do not establish risk.
- Seek a concrete alternative explanation or counterevidence before recommending intervention. Check expected cadence, comparable denominators, planned changes and telemetry completeness when interpreting usage.
- Separate supported concerns from consequential unknowns. Identify the smallest evidence request that could change the next action. A confirmed critical blocker is not cancelled by unrelated positive signals; repeated reports of one incident are not separate risk mechanisms.
- If a company policy is supplied, check its product, segment, lifecycle, timing and other stated applicability conditions. Cite the matching policy provision for any policy-specific recommendation. If scope mismatches or is unknown, do not apply it; continue the qualitative review and explain the limit. Do not invent thresholds, routing rules or SLAs.
- Map findings to proportionate proposals. Distinguish intervention, internal evidence verification and no intervention on current evidence. Use proposed owner roles and an evidence-based urgency rationale; never invent an assigned person, commitment or deadline. All external actions remain drafts.
Output
1. Findings
- Scope: account, as-of date, evidence reviewed and material coverage limits.
- Supported concerns: mechanism, supporting source references and bounded interpretation.
- Alternatives and unknowns: counterevidence, what the evidence does not establish and the most consequential gap.
- Reversal conditions: what new evidence would change each conclusion.
2. Action mapping
For each proposal, give the linked finding or unknown, next action, proposed owner role, timing rationale and prerequisites. Name any supplied policy/context that informed the recommendation. If none applies, say that routing and timing are proposals, not company policy.
A finding of no supported intervention is allowed; it is not an overall declaration of account health. Do not emit a composite health score, churn probability or claim that a review executed policy checks.
Completion checks
Every consequential claim is traceable; findings and hypotheses stay distinct; unknowns do not become healthy; proposed owners/dates do not become commitments; and the next action is proportionate to the evidence.
Examples and evaluation
See the illustrative walkthrough, evaluation cases and trigger cases.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

