基于 SOC 职业分类
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/benjam3n/reasoningtool --skill tr-template-registry命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
Route any input through a branching question tree to narrow down the optimal response strategy before writing. Two stages — PERCEIVE (classify input) then ACT (select response). Covers all prompt types.
Generate exhaustive guesses about user input using ALL search methods with coverage tracking. Guessing is SEARCH through possibility space. Tracks space created vs space covered to ensure comprehensive exploration.
Systematically evaluate and select from a set of guesses, options, or possibilities. Combines ARAW analysis with prioritization to determine which guesses are strong, weak, actionable, or eliminable.
| name | tr - Template Registry |
| description | Registry for domain templates. Store, retrieve, search, and manage domain-specific configurations. |
| output | {"format":"prose"} |
Input: $ARGUMENTS
Before executing, identify which interpretation matches the user's input:
Interpretation 1 — Performing a registry operation: The user wants to store, retrieve, search, list, compare, update, or delete a specific template in the registry. Interpretation 2 — Finding the right template for a task: The user has a problem or domain and wants the registry to match it to the best available template, so they can apply a pre-built analytical workflow. Interpretation 3 — Managing the registry itself: The user wants to audit, reorganize, or maintain the template collection (checking for overlaps, cleaning stale entries, validating integrity).
If ambiguous, ask: "I can help with a specific registry operation, finding a template for your task, or managing the registry — which fits?" If clear from context, proceed with the matching interpretation.
Manage domain templates:
REGISTRY STORE: [template name]
Template:
[Full template content from /dmt]
VALIDATION:
- [ ] Has required sections (description, triggers, skill_chain, dimensions)
- [ ] Triggers don't conflict with existing templates
- [ ] Skill chain references valid skills
- [ ] Version number present
STORE RESULT:
Status: [STORED / VALIDATION_FAILED / CONFLICT]
Location: templates/[domain_name].yaml
Conflicts: [any trigger conflicts with existing templates]
REGISTRY RETRIEVE: [template name]
SEARCH:
Looking for template: [name]
RESULT:
Status: [FOUND / NOT_FOUND / MULTIPLE_MATCHES]
If FOUND:
[Full template content]
If MULTIPLE_MATCHES:
Candidates:
1. [Template 1] - [brief description]
2. [Template 2] - [brief description]
Select one to retrieve.
REGISTRY MATCH: [user input]
INPUT: "[user's input text]"
TRIGGER SCAN:
| Template | Trigger Match | Score |
|----------|---------------|-------|
| [Template 1] | [matching trigger] | [0-100] |
| [Template 2] | [matching trigger] | [0-100] |
| [Template 3] | [no match] | 0 |
BEST MATCH:
Template: [name]
Trigger: [which trigger matched]
Confidence: [HIGH/MEDIUM/LOW]
RECOMMENDATION:
[Apply template X / No clear match - ask user / Multiple candidates]
REGISTRY LIST
AVAILABLE TEMPLATES:
| # | Template | Domain | Skills | Last Updated |
|---|----------|--------|--------|--------------|
| 1 | [name] | [domain] | [count] | [date] |
| 2 | [name] | [domain] | [count] | [date] |
| 3 | [name] | [domain] | [count] | [date] |
TOTAL: [N] templates
BY CATEGORY:
- Research: [list]
- Business: [list]
- Technical: [list]
- Creative: [list]
- Other: [list]
REGISTRY SEARCH: [keyword]
SEARCHING FOR: "[keyword]"
MATCHES:
In triggers:
- [Template 1]: trigger "[matching trigger]"
- [Template 2]: trigger "[matching trigger]"
In dimensions:
- [Template 3]: dimension "[matching dimension]"
In description:
- [Template 4]: "[matching text snippet]"
In assumptions:
- [Template 5]: assumption "[matching assumption]"
TOTAL MATCHES: [N]
MOST RELEVANT:
1. [Template] - [why relevant]
2. [Template] - [why relevant]
REGISTRY COMPARE: [template A] vs [template B]
COMPARING:
- Template A: [name]
- Template B: [name]
OVERLAP ANALYSIS:
Triggers:
| Template A | Template B | Overlap? |
|------------|------------|----------|
| [trigger] | [trigger] | [Yes/No] |
Dimensions:
| Dimension | In A? | In B? | Shared? |
|-----------|-------|-------|---------|
| [dim] | [Y/N] | [Y/N] | [Y/N] |
Skills:
| Skill | In A? | In B? | Order |
|-------|-------|-------|-------|
| [skill] | [Y/N] | [Y/N] | [A:1, B:2] |
Assumptions:
- Shared: [list]
- Only in A: [list]
- Only in B: [list]
OVERLAP SCORE: [0-100%]
RECOMMENDATION:
[Merge templates / Keep separate / One subsumes other]
REGISTRY DELETE: [template name]
DELETING: [template name]
PRE-DELETE CHECK:
- Template exists: [Yes/No]
- Dependencies: [list any templates that reference this one]
- Usage stats: [how often used, if tracked]
CONFIRMATION REQUIRED:
Type "CONFIRM DELETE [template name]" to proceed.
DELETE RESULT:
Status: [DELETED / CANCELLED / NOT_FOUND]
REGISTRY UPDATE: [template name]
CURRENT VERSION: [version number]
CHANGES:
[Specify what to change]
UPDATED FIELDS:
- [Field 1]: [old value] -> [new value]
- [Field 2]: [old value] -> [new value]
NEW VERSION: [incremented version]
VALIDATION:
- [ ] Changes don't break skill chain
- [ ] Triggers still unique
- [ ] Version incremented
UPDATE RESULT:
Status: [UPDATED / VALIDATION_FAILED]
Templates are stored as YAML files:
templates/
|-- index.yaml # Registry index
|-- research/
| |-- academic.yaml
| |-- market.yaml
|-- business/
| |-- consulting.yaml
| |-- strategy.yaml
|-- technical/
| |-- software.yaml
| |-- data_science.yaml
|-- creative/
|-- writing.yaml
|-- design.yaml
# templates/index.yaml
version: 1.0
last_updated: [date]
total_templates: [N]
templates:
- name: academic_research
path: research/academic.yaml
triggers: [research, literature, hypothesis]
version: 1.2
- name: consulting
path: business/consulting.yaml
triggers: [client, engagement, deliverable]
version: 1.0
trigger_map:
research: academic_research
literature: academic_research
hypothesis: academic_research
client: consulting
engagement: consulting
QUICK MATCH: [input]
Best template: [name] ([confidence]%)
Apply with: /apply_template [name]
QUICK LIST
Templates: [name1], [name2], [name3], ...
Total: [N]
User: I need to do a literature review on machine learning fairness
REGISTRY MATCH: "literature review on machine learning fairness"
TRIGGER SCAN:
| Template | Trigger Match | Score |
|----------|---------------|-------|
| academic_research | "literature review" | 95 |
| data_science | "machine learning" | 60 |
| technical_writing | no match | 0 |
BEST MATCH:
Template: academic_research
Trigger: "literature review"
Confidence: HIGH
RECOMMENDATION:
Apply academic_research template. This will:
1. Extract research assumptions
2. Discover relevant dimensions (methodology, scope, etc.)
3. Enumerate the research space
4. Find methodological analogies
5. Validate MECE coverage
6. Synthesize insights
Proceed? [Yes/No/Show template first]
Before completing registry operation:
Use with:
/dmt - Create templates to store/tpm - Keep templates updated