Influencer Discovery
Build and screen a tiered creator shortlist from scratch for a defined brand, niche, or platform.
The skill limits itself to user-provided or public data, defaults to keyless operation, and separates roster proposals from canonical registry writes. However, it writes memory, promotes items to HOT cache, and submits registry proposals without detailed per-write confirmation, rollback, retention, or sensitive-contact handling rules; publisher identity is also unverified, so points are deducted.
Inputs, outputs, procedures, templates, and uncertainty handling such as `unconfirmed` are reasonably specified, with connector-free degradation. Still, the workflow depends on several unprovided reference files and external platforms, and no skill-specific key-path tests or diagnosable failure feedback are shown. Static calibration therefore keeps the score conservative.
Triggers, audience, required parameters, and the boundary against ranking a known shortlist are clear. Evidence for cross-platform availability, mainland-China reachability, and Chinese-market operating alternatives is limited; `geo-relevance` is low and the skill does not provide a dedicated China-compatible path, so points are deducted.
The skill has a structured contract, quick starts, templates, an example, version and Apache-2.0 metadata, related-skill links, and repository security documentation. Maintenance ownership, changelog/update path, and single-skill dependency notes are less explicit, and the skill's 18.0.0 version conflicts with the security policy's supported 17.x line, so full marks are not justified.
The intended report is directly structured around a candidate pool, profiles, screening, tiers, and handoff, and can operate on pasted inputs. But discovery and metric verification largely remain manual, live retrieval is optional, and the worked-example counts and scores lack verifiable execution evidence; substantial human validation remains necessary, so the static score is limited.
The method asks users to log queries, assess source quality, and mark unconfirmed facts instead of guessing. However, there are no skill-specific tests, third-party execution records, or corroborating evidence sets in the supplied files; the evidence is primarily procedural and illustrative, so the score is deducted.
- Core discovery depends on user-supplied public material or reachable external platforms; the example counts, engagement rates, and fit scores should not be treated as verified facts.
- Obtain explicit confirmation before writing memory, HOT cache, or creator-registry proposals, and review minimization, retention, and deletion handling for contact data.
- For mainland-China use, separately verify platform reachability, data completeness, and compliance boundaries; unavailable data should remain marked unverified rather than inferred.
What does this skill do, and when should you use it?
Influencer Discovery is the scout-phase skill in Aaron Marketing Skills for building creator candidate pools around a brand, niche, platform, or follower range. It captures search criteria, searches through hashtags, similar accounts, competitor mentions, and native platform discovery, then screens candidates for audience fit, engagement, recency, relevance, and brand-safety risks. It produces candidate statistics, creator profiles, preliminary discovery-triage signals, and a three-tier shortlist. It does not calculate STAR Suitability; that downstream read belongs to fit-scorer.
Reads the brand or product, niche, platforms, follower range, engagement floor, location or language, audience demographics, exclusions, available brand and audience context, and existing creator records. It logs discovery queries, screens for follower and engagement requirements, relevance, recency, suspected fake followers, controversy, competitor exclusivity, and inactivity, then builds profiles covering basics, metrics, audience, content, partnership history, contact paths, and preliminary triage signals. It compiles must-reach, strong, and consider tiers with next steps. Keyless YouTube, TikTok, and X oEmbed metadata can populate profile fields; a provided Python YouTube connector can retrieve displayed subscribers, total views, video count, and recent video metrics.
- A brand entering a new niche needs a creator roster built from scratch.
- A marketing team is expanding to a new platform and wants micro- or nano-creators within a specified follower and engagement band.
- A team needs replacement partners after creator churn and wants an initial authenticity and brand-safety screen.
- A brand wants to identify creators associated with competitors before evaluating partnership options.
- An existing roster must be deduplicated and enriched before being handed to fit-scorer for weighted ranking.
What are this skill's strengths and limitations?
- Covers criteria definition, discovery, initial screening, creator profiling, tiering, and handoff in one focused workflow.
- Separates preliminary discovery signals from STAR Suitability scoring and verified STAR veto decisions.
- Works without required live integrations and supports user-supplied inputs plus keyless metadata paths.
- Defines deduplication and a clear downstream handoff to fit-scorer.
- Follower and engagement data generally remain user-supplied, manually exported, or connector-dependent; oEmbed does not provide those metrics.
- It does not perform STAR scoring, final veto judgments, outreach, contracting, or performance reporting.
- Memory saves, shortlist caching, and registry proposals require separate authorization and are not automatic.
- The supplied material does not show an independent test suite or evidence of complete platform coverage for this individual skill.
How do you install this skill?
For Claude Code, run /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, then /plugin install aaron-marketing@aaron. For other Agent Skills-compatible hosts, run npx skills add aaron-he-zhu/aaron-marketing-skills. Alternatively, run git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. The supplied README does not document a standalone copy command for this skill directory.
How do you use this skill?
State the brand or niche and provide platforms, follower range, engagement floor, location or language, audience requirements, and exclusions where possible. Example: Find influencers in sustainable fashion with 10K-100K followers for my eco clothing brand. If required criteria are missing, the skill should return NEEDS_INPUT rather than fabricate candidates. The report is returned inline by default; saving, caching, and creator-registry proposals require separate exact authorization.
How does this skill compare with similar options?
fit-scorer is the next-stage alternative for weighted STAR Suitability scoring and ranking of a known shortlist. audience-mapper is used when the target audience is still unclear, while competitor-tracker maps the competitive creator field.