Skip to content
SkillModel-led309 tokens in use

ads-research

Refreshes the platform, API, policy and benchmark evidence behind Claude Ads, and demotes any claim that cannot be reverified from a current source.

11,191 stars on the repository MIT LicenseAgriciDaniel/claude-ads · skills/ads-research
/plugin marketplace add agricidaniel/claude-ads /plugin install claude-ads@ai-marketing-hub-claude-ads
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing ads-research costs

70tokens idle
309tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter8/15
  • Structure12/25
  • Examples0/15
  • Output5/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity84
  • Maintenance100
  • Community56
  • Maturity32
  • Momentum40

ads-research keeps the knowledge base of the Claude Ads toolkit current. It selects the overdue or requested slice of evidence, sends bounded research workers out by platform or topic, and records each source with its publisher, dates, geography, confidence and license.

A separate verifier checks the claims that matter most. When a claim has passed its refresh date and cannot be reverified, the skill demotes it for the current run and blocks any release-current assertion that depends on it.

What You Get

  • Refreshed evidence records with URL, publisher, publication and retrieval dates, geography, availability and confidence
  • A separate verifier pass on load-bearing claims
  • A list of contradictions and demoted claims
  • Proposed canonical updates with affected controls, skills, adapters and tests
  • A stated blocker and recovery path when a source or tool is unavailable

Common Use Cases

  • Running a refresh after a platform changes its API or policy
  • Checking which claims passed their refresh date before a release
  • Reviewing ecosystem issues, pull requests and forks for relevant changes
  • Updating benchmark evidence for a platform audit

What Makes It Unique

An expired refresh date means the claim is not current, and the skill says so out loud. If official sources cannot be reached, the claim is demoted for the run and dependent release-current assertions are blocked, with the missing access named as the recovery step. Findings from a worker never become canonical on their own.

Under The Hood

The skill is written instructions with no scripts. The agent reads the repository’s control-plane contracts, dispatches research workers that fetch web pages and read official sources, and sends a separate verifier agent to check load-bearing claims. Fetched content is treated as untrusted data, and restricted prompts, large source passages, unlicensed code, issue text and private account material are kept out of the repository. Updates to the repository’s evidence files are proposed for review.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers