메뉴
Build and maintain structured, persistent knowledge graphs across sessions. Knowledge-ops transforms ephemeral session findings into reusable intelligence — connecting entities, tracking confidence over time, and enabling recall across engagements.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/brucesongs/kali-claw --skill knowledge-ops명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | knowledge-ops |
| description | Build and maintain structured, persistent knowledge graphs across sessions. Knowledge-ops transforms ephemeral session findings into reusable intelligence — connecting entities, tracking confidence over time, and enabling recall across engagements. |
| origin | openclaw |
| version | 0.2.0.2 |
| compatibility | ["openclaw","claude-code","cursor","windsurf"] |
| allowed-tools | ["Bash","Read","Write","Edit","WebSearch","WebFetch","Agent"] |
| defense_triple_required | false |
| metadata | {"domain":"knowledge","tool_count":0,"guide_count":5,"last_reviewed":"2026-07-26"} |
Knowledge-ops transforms ephemeral session findings into reusable intelligence — connecting entities, tracking confidence over time, and enabling recall across engagements.
Domain: knowledge
| Attribute | Value |
|---|---|
| Domain | Knowledge Operations |
| Skill ID | knowledge-ops |
| Version | 1.0.0 |
| Hacker Laws | Law 3 (Intelligence Over Force), Law 8 (Learn from Every Operation), Law 9 (Systematic Over Random) |
| Related Skills | codebase-onboarding, deep-research, social-intelligence |
Build and maintain structured, persistent knowledge graphs across sessions. Knowledge-ops transforms ephemeral session findings into reusable intelligence — connecting entities, tracking confidence over time, and enabling recall across engagements.
Without knowledge-ops, every session starts from zero. With it, each session builds on the last.
The atomic element of the knowledge graph. Each unit contains:
{
"id": "KU-2026-05-001",
"type": "finding | entity | relationship | pattern | hypothesis",
"created": "2026-05-11",
"updated": "2026-05-11",
"confidence": 78,
"tags": ["auth", "JWT",
A network of linked knowledge units. Relationships between units reveal:
A collection of knowledge graphs organized by domain (target, technology, vulnerability type). Stored in the memory/ directory as markdown files with structured frontmatter.
| Type | Description | Example |
|---|---|---|
entity | A person, system, domain, IP, org | target.com, admin@target.com, 192.168.1.1 |
finding | A discovered fact about a target | SQL injection in search endpoint |
relationship | Link between two entities | User X administers System Y |
pattern | Recurring vulnerability or behavior | This org consistently uses outdated JWT libs |
hypothesis | Unconfirmed, needs validation | Admin portal may be accessible via VPN bypass |
intelligence | Aggregated, analyzed insight | Target org has poor patch management practices |
Extract knowledge units from session outputs:
Assign type, tags, confidence, and links to each unit. Establish relationships between related units.
Write structured units to the appropriate memory file. Follow the naming convention: memory/YYYY-MM-DD-[topic].md.
At session start, query the knowledge base for relevant units. Use tags, entity names, and linked IDs to surface prior context.
Update confidence scores as findings are confirmed or refuted. Mark expired intelligence. Remove stale hypotheses that were disproven.
Confidence scores represent certainty about a knowledge unit:
| Score | Meaning | Action |
|---|---|---|
| 0–25 | Speculation — no evidence | Label as hypothesis |
| 26–50 | Unconfirmed — single source | Needs validation |
| 51–75 | Probable — multiple consistent signals | Use with caution |
| 76–90 | High confidence — independently confirmed | Act on it |
| 91–100 | Verified — demonstrated/exploited | Treat as fact |
Confidence is not static. Update it when:
---
id: KU-[YYYY]-[MM]-[NNN]
type: [finding|entity|relationship|pattern|hypothesis|intelligence]
target: [target name or scope]
confidence: [0-100]
created: [YYYY-MM-DD]
updated: [YYYY-MM-DD]
tags: [tag1, tag2, tag3]
linked: [KU-ID-1, KU-ID-2]
source: [skill that produced this]
expires: [YYYY-MM-DD or null]
---
## Detail
[Full context, evidence, commands used, observed behavior]
## Connections
- **Confirms**: [what this validates]
- **Contradicts**: [what this disproves]
- **Leads to**: [next investigation or action]
## Confidence History
| Date | Score | Reason |
|------|-------|--------|
| [date] | [score] | [why this score] |
index=kb sourcetype=document_access | stats dc(doc_id) by user_id | where dc > 100