Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Source Verification — measured AI agent skill | DecimalAI

nwave-ai/nw-source-verification

NewSkillScoreMaintained

Source reputation tiers, cross-referencing methodology, bias detection, and citation format requirements

researchby nwave-ai

Links this skill to your agents and keeps it current with the author’s updates.

On-demandOn-demand: invoked explicitly — never occupies your agent's context until called.

Research pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/nwave-ai-nw-source-verification/SKILL.md

decimalai skills pull nwave-ai-nw-source-verification --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/nwave-ai-nw-source-verification/SKILL.md https://app.decimal.ai/s/nwave-ai-nw-source-verification/SKILL.md

Raw file: /s/nwave-ai-nw-source-verification/SKILL.md ↗ · machine-readable summary at /s/nwave-ai-nw-source-verification.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 3h ago

Yes — +50% lift+4% tokens0% turns95% of whole cases pass with skill (21/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-02	✗→✓	▲ Improved	-13%	0%
case-03	✗→✓	▲ Improved	-12%	0%
case-04	✗→✓	▲ Improved	-40%	0%
case-01	✗→✓	▲ Improved	128%	0%
case-05	✗→✓	▲ Improved	-19%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

nwave-ai / nwave

nWave/skills/nw-source-verification/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1

Skill contents

SKILL.md

Source Verification

Source Reputation Tiers

Validate every source against the trusted source domains provided via prompt context.

| Tier | Score | Examples | Verification | |------|-------|---------|-------------| | High | 1.0 | Academic (.edu, arxiv.org, ieee.org), Official (.gov, w3.org, ietf.org), Tech docs (developer.mozilla.org), OSS foundations (apache.org, cncf.io) | Standard citation | | Medium-High | 0.8 | Industry leaders (martinfowler.com, stackoverflow.com, infoq.com) | Cross-ref with 1+ high-tier | | Medium | 0.6 | Community (medium.com verified experts, dev.to, hashnode.com) | Author verification + 3-source cross-ref | | Excluded | 0.0 | Unverified blogs (.blogspot.com, wordpress.com), quora.com, pastebin.com | Reject, log warning, find alternative |

Cross-Referencing Methodology

Identify the specific assertion to verify
Find 2+ independent sources not citing each other (avoid circular refs)
Verify independence: different authors, publishers, organizations
Compare: agree on substance (minor wording differences OK)
Document: verified / partially verified / unverified per finding

Circular Reference Detection

Source B cites Source A = one source, not two
Multiple sources referencing single study = cite the original
Prefer primary over secondary sources

Bias Detection Checklist

Evaluate before citing:

Commercial interest: selling related product/service?
Sponsorship: sponsored/funded content?
Conflict of interest: author benefits from conclusion?
Geographic/cultural bias: limited to single region?
Temporal bias: publication dates skewed to specific era?
Cherry-picking: contradictory evidence acknowledged?
Logical fallacies: correlation as causation, authority without evidence

When bias detected: note in Source Analysis, reduce confidence.

Citation Format

[1] {Author/Organization}. "{Title}". {Publication/Website}. {Date}. {Full URL}. Accessed {YYYY-MM-DD}.

Required Metadata Per Source

Source URL | Domain | Access date | Reputation score (from tiers) | Verification status

Paywalled or Restricted Sources

Mark "Paywalled]"/"Restricted Access]" | Provide URL | Find open-access alternative | Note in Knowledge Gaps

Want to see if nwave-ai/nw-source-verification would activate on your phrasing? Try it before you install — no LLM call, no signup.

Other measured skills in the registry, with their headline benchmark lift.