Clawdna

Generate a public, privacy-safe persona/card/wiki lead from historical behavior when the user explicitly asks to run ClawDNA.

Security Scan

Instruction-only skill whose stated purpose (generate a redacted public persona from historical activity) matches its requested scope, but it relies on the agent/platform to enforce explicit consent and redaction rules — verify consent prompts and test with non-sensitive data first.

Purpose & Capability

Name and description (generate privacy-safe persona from historical behavior) align with the instructions: the SKILL.md explains how to index, chunk, extract repeated patterns, redact details, and produce a fixed output template. The listed input locations (agent session files and memory files) are consistent with the stated goal of using historical agent activity.

Instruction Scope

The instructions explicitly permit reading local agent session files (~/.openclaw/agents/<runtime-agent-id>/sessions/*.jsonl) and memory files (memory/*.md, MEMORY.md) but only after explicit user confirmation; they forbid auto-trigger beyond explicit user commands and require redaction and no raw transcript excerpts. This is coherent, but the safety depends on the platform actually enforcing the consent step. Also, the skill directs no audit appendix be returned, which reduces transparency about what was redacted.

Install Mechanism

No install spec or code files; instruction-only skills are lowest-risk for on-disk installs. Nothing will be downloaded or executed by an installer from this package itself.

Credentials

The skill requests no environment variables, credentials, or unusual config paths beyond the allowed-in-scope local files explicitly documented in SKILL.md. The requested data sources are proportionate to the stated purpose.

Persistence & Privilege

always is false and the skill is user-invocable only — appropriate. However, the SKILL.md's reliance on explicit trigger/consent is an instruction-level control; actual enforcement depends on the host platform's trigger & consent UX. Autonomous invocation is platform-default but not expressed as required by the skill.

Assessment

This skill appears internally consistent with its goal, but it depends on the platform and agent to prompt for and enforce explicit consent before reading local session or memory files. Before installing or using it: (1) verify your platform will prompt you and require explicit approval each time the skill wants to read ~/.openclaw/... or memory files; (2) test the skill on non-sensitive/example data to confirm redaction behavior and that no raw transcripts or sensitive excerpts are returned; (3) be aware the SKILL.md asks for no audit appendix, so you may not get a detailed disclosure of what was redacted — ask for an explicit confirmation message if you need transparency; (4) note a minor metadata mismatch (registry shows v0.1.4 while _meta.json inside is v0.1.2) — prefer the latest published source or ask the author for clarification. If you need stronger guarantees (e.g., proof of redaction or local-only processing), seek a skill that provides verifiable audit logs or runs locally-only with explicit UI consent flows.

Goal

Create a concise public profile based on real historical behavior, not guesses.

Trigger

Run only on explicit user intent, such as:

• use ClawDNA
• /clawdna
• generate with ClawDNA

Do not auto-trigger from generic chat.

Safety Mode (Default)

Default to minimal analysis scope.

• Prefer user-provided summaries/history exports when available.
• If local history access is needed, request explicit confirmation first.

Allowed Data Scope

Only these locations are in scope:

1. ~/.openclaw/agents/<runtime-agent-id>/sessions/*.jsonl
2. memory/*.md
3. MEMORY.md (only if allowed in current context)

Never auto-expand to other directories. Never access other agents' logs without explicit user consent.

Non-Negotiable Rules

1. No speculation; omit unsupported claims.
2. Public privacy output only (redacted by default).
3. Show time span only in coverage note (no file count).
4. Output language follows user language.
5. Proper nouns/names stay in original form.
6. Core Capabilities and Representative Work are title-only.
7. Data minimization: extract only fields needed for profile.
8. Do not output raw transcript excerpts.

Large-History Strategy

If history is too large:

1. Build metadata index first.
2. Process by time chunks.
3. Summarize each chunk with fixed fields.
4. Merge summaries and keep cross-window repeated patterns.

Extraction Method

1. Build timeline bounds (first/last active).
2. Detect recurring behavior signals across windows:
   • execution
   • collaboration
   • analysis
   • operations
   • creation
   • governance/boundaries
3. Keep only stable repeated patterns.
4. Redact sensitive details.
5. Map to fixed output template.

Output

Return one Markdown display profile only. No audit appendix.

Fixed Output Template

{{Name}}

{{LeadLabel}}

{{identity + start time + stable traits + value}}

{{InfoboxLabel}}

• {{NameLabel}}: {{}}
• {{TypeLabel}}: {{execution/creative/analysis/operations/hybrid}}
• {{FirstActivationLabel}}: {{YYYY-MM-DD}}
• {{ActiveTimeSpanLabel}}: {{Start ~ Now}}
• {{TotalTokensLabel}}: {{optional, only if reliably available}}
• {{PrimaryDomainsLabel}}: {{max 3}}
• {{InteractionStyleLabel}}: {{}}
• {{CollaborationModeLabel}}: {{}}
• {{DefaultPrinciplesLabel}}: {{privacy-first / risk-confirmation / rollback-first}}

{{OriginEvolutionLabel}}

• {{InitialStageLabel}}: {{}}
• {{EvolutionStageLabel}}: {{}}
• {{CurrentStageLabel}}: {{}}

{{OperatingMethodLabel}}

• {{}}
• {{}}
• {{}}

{{PersonalitySnapshotLabel}}

{{2-3 behavior-based sentences}}

{{CoreCapabilitiesLabel}}

• {{title only}}
• {{title only}}
• {{title only}}

{{RepresentativeWorkLabel}}

• {{title only}}
• {{title only}}
• {{title only}}

{{MilestonesLabel}}

• {{Date}}: {{}}
• {{Date}}: {{}}
• {{Date}}: {{}}

{{CollaborationGuideLabel}}

• {{BestInputLabel}}: {{goal / constraints / priority / deadline}}
• {{BestRhythmLabel}}: {{}}
• {{PreferredOutputLabel}}: {{}}

{{BoundariesSafetyLabel}}

• {{}}
• {{}}
• {{}}

{{PersonaTagsLabel}}

{{tag1}} / {{tag2}} / {{tag3}} / {{tag4}}

{{EndingLabel}}

Preset persona = initial intent; long-term behavior = real persona; user habits = persona shaper.

Final Checks

• Explicit trigger confirmed
• Scope consent confirmed if local logs used
• No speculative claims
• Title-only for capabilities/work
• Privacy-safe wording