2.6k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
3 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill clawdna- _meta.json624 B
- SKILL.md3.5 KB
Overview
This skill generates a concise, privacy-safe, wiki-style profile from an agent's OpenClaw history when you explicitly ask to run ClawDNA. It produces factual, standardized summaries that highlight stable behaviors, capabilities, and representative work while redacting sensitive data. The output is designed for public sharing and excludes raw transcripts or secrets.
How this skill works
On explicit user command, the skill indexes allowed OpenClaw session paths and any permitted memory files, building a metadata-first view to avoid loading raw logs at once. It uses jq for structured extraction and falls back to asking the user to install jq if missing. The tool processes history in time-chunked summaries, filters for repeated cross-window patterns, applies privacy redaction, and maps stable signals into a fixed wiki-style template.
When to use it
- When you explicitly request 'use ClawDNA', '/clawdna', or 'generate with ClawDNA'.
- When you want a public-facing, privacy-safe profile derived from full OpenClaw history.
- When you consent to full-history analysis in the current request rather than a recent-only window.
- When you need a standardized summary for archival, documentation, or public presentation.
Best practices
- Always include explicit consent in the same request to enable full-history mode.
- Ensure jq is installed before running; pause and install jq if it's not available.
- Limit scope to the current runtime agent and allowed paths unless you explicitly request others.
- Confirm you are comfortable with time-span disclosure but not file counts or raw excerpts.
- Provide goal, constraints, and desired granularity to guide extraction and summary style.
Example use cases
- Create a public profile for an agent to publish on an archive or documentation site.
- Generate an anonymized, standardized summary for a team retrospective or handoff.
- Produce a concise capabilities overview to include in a project catalog or README.
- Audit recurring behaviors and representative work while ensuring privacy-first redaction.
FAQ
It only reads allowed paths for the current runtime agent and any explicitly permitted memory files; it will not expand to other directories without your confirmation.
Is sensitive data kept or shown?
No. The skill applies privacy redaction, removes secrets and direct identifiers, and never outputs raw transcript excerpts.
Can I run this without jq?
No. jq is required. The skill will stop and ask you to install jq before proceeding.