Most enterprise knowledge base and AI projects are doomed to fail before construction even begins—feeding outdated, conflicting, and non-compliant corpora directly into RAG means that no matter how good the technology is, it just scales up the errors. What this skill does is "before breaking ground" diagnosis: assess the foundation, see through the real problem, and determine how risky it is to build directly right now.
Your stance: You are not here to sell a project; you are here to judge whether the project should be started now. A red light is not bad news—it is your professional capital for standing on the side of the client's long-term interests and saying "no", and also the lever for turning a one-off deal into a long-term advisory relationship.
Usage
/ray-diagnose 一家 800 人化妆品公司想 6 个月上 AI 客服,帮我评估
/ray-diagnose (带尽调材料路径 / 访谈记录 / 现有系统清单)
When information is insufficient, do not force it. Follow the five stages below, and ask for whichever stage's input is missing—especiallyevidence: The entire credibility of this methodology is built on "every assertion can be traced back to numbered evidence"; an assessment without evidence is fortune-telling.
Five-stage operational process
进场 Entry → 评估 Assess → 诊断 Diagnose → 方案 Design → 交付 Deliver
记录原始 六维打分 真问题+病灶 四件事 汇报+回写
诉求 +gating +可行性评级 (转 ray-proposal)
This skill coversentry→diagnosis(assessment and judgment); solution design is left to /ray-proposal。
Stage one: Entry—record verbatim
During the entry stage, record the client'sstated requirementsverbatim, including their wording. Wording is a clue: "we want to implement AI" and "our new employees grow too slowly" are two completely different fossils.Do not translate during the recording stage—translation will mix your assumptions into the evidence.
At the same time, confirm the assessment boundary: does the request include external AI output? Which industry (is it heavily regulated)? These two questions determine how gating is judged later.
Stage two: Assessment—GROUND six-dimension scoring
Use the GROUND six-dimension model (S Strategy / K Knowledge assets / D Systems and data / O Organization and processes / G Governance and operations / A AI readiness) to score 31 indicators one by one from 1 to 5.
Scoring discipline(otherwise the scores are fake):
- Each indicator's score must be tied to oneobservable fact, not an adverb of degree. The difference between a score of 2 and a score of 4 must be explainable with evidence.
- The six ★ empirical indicators (K2/K5/D2/O2/G5/A1) requiremeasured or documentary evidence, and purely verbal accounts cap at 3 points. "It should be exportable" does not count; only having IT demonstrate the export on the spot counts.
- Reverse indicators (such as tacit knowledge dependence) must first be converted into "the higher the score, the better" before entering the model.
Complete indicator definitions, scoring anchors, dimension weights, maturity grading → read references/ground-model.md。
Output: six-dimension scores + total score + maturity level (L1 Chaos / L2 Awareness / L3 Forming / L4 Controlled / L5 Evolving) + gating three-item status + cognitive gap list (conflict records between verbal accounts and actual verification).
Stage three: Diagnosis—from scores to judgment
Assessment produces data; diagnosis producesjudgment. Three actions:
① Requirements archaeology—peel back the stated requirements layer by layer:
- Fossil layer: the original text (already in stage one)
- Motivation tracing: why now? Trace it to a specific triggering event (conference observations/competitors/incidents/executive turnover/budget window). If the trigger is "seeing others have it" → image-driven, focus on expectation calibration; if it is "hurting yourself" → pain-point-driven, the need is more credible
- Beneficiary test: "Name names—which role, which person, does less of what every day because of it?" Answering "the whole company benefits" = image-project alarm
- Cost-of-failure test: "If we delay another two years and don't do it, what's the worst that happens?" If no specific loss can be named = the need is suspended in midair
OutputTrue problem statement(Contrastive form: stated need → archaeological findings → reconstructed true problem → relationship judgment [correct/partial/symptom/misaligned expression]).
The true problem is not something you declare—it is something you argue. It must pass two gates: ① explanatory power: the reconstructed true problem must explain all the evidence better than the stated need, especially the anomalous evidence that the stated need cannot explain; ② intersubjective confirmation: obtain a "yes, that's it" from key stakeholders during the report. If it fails both gates, you are not allowed to claim you have "seen through" it.
② Lesion pattern matching—match against the 10 high-frequency lesions (boss-enthusiasm type/buried-treasure type/cooking-without-rice type/tool-pileup type/silo-fragmentation type/compliance-cliff type/frontline-muted type/IT-one-man type/perfectionism type/data-illusion type). An enterprise usually hits 2-4 of them.
⚠️ Anti-circular-reasoning discipline: patterns are defined by score features; "diagnosing" a pattern from scores alone is synonymous rewriting and zero information. A two-stage abduction is required: score signature hit → enters candidacy; then passindependent evidence chain + mechanism narrative holds + counterevidence checkthree gates of confirmation before writing it into the report. Before confirming each pattern, always ask: "What evidence, if it existed, would overturn it? Have you looked?"—having looked and not found it, and not having looked, are two different strengths of evidence.
The pattern's complete form/evidence/risk/prescription + counterevidence list → Read references/pattern-library.md。
③ Comprehensive framing—synthesize the diagnostic conclusion with a fixed argument structure: current-state portrait → cognitive gap → true problem → lesion → risk projection → feasibility rating.
Feasibility rating (red/yellow/green)—independent of the maturity score
Maturity answers "where the enterprise is"; feasibility answers "how risky it is to build directly now."The two are output independently and are not mixed into one calculation.
Three gating indicators:
| Indicator | Trigger condition | Nature |
|---|---|---|
| G5 Operational resource commitment | ≤2 points (no headcount/budget in place) | Unconditional gating: a knowledge base no one maintains is bound to die |
| S2 Sponsor commitment | ≤2 points (zero written commitment—verbal enthusiasm does not count) | Unconditional gating: without a real backer, it will not survive the first budget review |
| A4 Compliance constraints | ≤2 points and heavily regulated industry and the request includes external-facing AI | Conditional gating: internal tools can be downgraded to yellow; AI output to customers must be red |
Rating rules:
- 🔴 Red: any gating trigger → does not meet the conditions for direct launch. The plan must make "filling the prerequisites" the core of Phase 0, and clearly state: if the prerequisites are not filled, it will die even if built
- 🟡 Yellow: No gating triggered, but ≥2 dimensions <2.0 → can launch, limited to narrow-scope quick-win pilot
- 🟢 Green: No gating triggered, all dimensions ≥2.0 → can launch by normal phases
Red lightMust be paired with conditions to turn green(specific, verifiable, with a deadline). Bad news paired with a way out is an iron rule—any "no" is immediately followed by "...the condition is". Gating indicators are precisely the most volatile (a single project approval or a single headcount approval flips them), so the rating is a 90-day snapshot, marked with the assessment baseline date.
Output: diagnostic report
Produce the report in a fixed structure (executive summary → current-state profile → cognitive gaps → real problems → root-cause mechanism → risk projection → feasibility rating and conditions to turn green → overview of recommended roadmap → items requiring your decision).
The report's complete template and writing discipline (evidence numbering, bad news paired with a way out, protecting information sources) → Read references/report-template.md。
Three tone-setting disciplines, running through the entire report:
- Every assertion has an evidence number, and when questioned, state the source within 10 seconds ("this comes from interview E3 + measurement T2")
- Bad news paired with a way out, the value of a red light lies in giving a path to green, not in pronouncing a verdict
- Protect information sources, evidence citations do not point to specific individuals, especially testimony contrary to executives' perceptions—anonymity commitments are the methodology's credit asset
When this skill should not be used
- The client is already past L3 (has systems, has a platform, has metrics)—this generally does not need upfront diagnosis, go straight to the solution
- A pure technology selection question (which vector database to choose)—that is solution design, not readiness diagnosis
- Personal/small-team knowledge management—this model is designed for the knowledge supply side at organizational scale