Install
$ agentstack add skill-chuongdlb-agent-skills-kb-integrator ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
KB Integrator — Novelty Classification and Topic Integration
Purpose
Take one paper card (output of paper-extractor) and integrate it into the knowledge base by:
- Classifying its novelty against existing KB content
- Updating relevant topic files
- Updating the index, timeline, and registry
This skill must run sequentially — it writes to shared topic files and the registry.
When to Use
Invoke this skill when:
- A new paper card exists in
kb/papers/withnovelty: pending - You are running a
kb-maintenancecycle - You are doing a cold start / bulk integration
Not for: Extracting papers (use paper-extractor), or querying the KB (use kb-query).
Input
- Paper card path:
kb/papers/.md - Existing KB:
kb/topics/,kb/index.md,kb/registry.json - Scoring rubric:
kb/config/scoring-rubric.md
Novelty Classification Algorithm
Step 1: Load Context
- Read the paper card's YAML frontmatter to get
domainsandtype - Read
kb/config/scoring-rubric.mdfor classification thresholds - Load relevant topic files based on domain tags (Glob
kb/topics/*.md, filter by domain match) - Load
kb/registry.jsonfor dedup check
Step 2: Dedup Check
Check kb/registry.json for:
- Matching DOI
- Matching arXiv ID
- Title similarity (exact lowercase match after stripping punctuation)
If duplicate found: skip integration, log to timeline as "Skipped (duplicate of )".
Step 3: Classify Novelty
For each contribution in the paper card:
- Scan existing topic files for SoTA entries on the same benchmark/task
- Compare results: If the paper's results table shows >10% improvement over the current SoTA entry →
sota - Check for new capability: If the paper demonstrates something no existing entry does →
sota - Check for incremental: If `
Step 5: Update Topic Files
For each domain tag in the paper card:
- If topic file doesn't exist → create it from the template below
- If paper is
sota:
- Add full write-up to the "State of the Art" section
- Move the previous SoTA entry to a "Previous SoTA" sub-section (keep 2-3 lines)
- Update the "Methods Landscape" table
- Add to "Papers Contributing" list
- If paper is
incremental:
- Add 2-3 line summary to "Papers Contributing" noting specific addition over SoTA
- Update "Methods Landscape" table if it adds a new method
- If paper is
complementary:
- Add to "Papers Contributing" with 3-5 line summary
- Update "Methods Landscape" table
- Consider adding to "Open Problems" if it identifies new gaps
- If paper is
derivative:
- Add 1-2 line entry to "Papers Contributing"
- If paper is
survey:
- Add entry to a "Surveys" section (create if needed)
- Check if survey identifies open problems not in "Open Problems" → add them
Step 6: Update Index
Add a row to kb/index.md:
| | | | | | | | [card](papers/.md) |
Step 7: Update Timeline
Add entry to kb/timeline.md:
## YYYY-MM-DD
- Added: — []
- Updated: —
Step 8: Update Registry
Add entry to kb/registry.json:
{
"id": "-",
"title": "",
"doi": "",
"arxiv": "",
"title_hash": "",
"domains": ["domain1", "domain2"],
"novelty": "",
"added": "YYYY-MM-DD"
}
Increment total_papers.
Step 9: Update Stats
Update kb/stats.md with new counts.
Topic File Template
When creating a new topic file:
#
## State of the Art
## Methods Landscape
| Method | Paper | Year | Approach | Key Strength | Key Limitation |
|--------|-------|------|----------|-------------|----------------|
## Open Problems
## Surveys
## Papers Contributing
| Paper | Year | Novelty | Summary |
|-------|------|---------|---------|
Quality Checks
After integration, verify:
- [ ] Paper card
noveltyfield updated frompending - [ ] At least one topic file updated
- [ ] Index has new row
- [ ] Timeline has new entry
- [ ] Registry has new entry with correct field values
- [ ] Registry
total_papersmatches actual count - [ ] No duplicate entries in registry
Handling Edge Cases
Paper spans multiple domains
Update ALL relevant topic files. The paper appears once in the index/registry but in multiple topic files.
No existing topic file matches
Create a new topic file. Use the domain tag as the filename slug (e.g., drone-simulation.md).
Paper supersedes existing SoTA
Move the old SoTA to "Previous SoTA" section. Keep 2-3 lines summarizing what it contributed. The new paper gets the full write-up.
Cold start (empty KB)
The first paper in each domain automatically gets sota classification since there's nothing to compare against. Mark it as sota (initial) in the topic file.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: chuongdlb
- Source: chuongdlb/agent-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.