ads-audit
Full multi-platform paid advertising audit with parallel subagent delegation. Analyzes Google Ads, Meta Ads, LinkedIn Ads, TikTok Ads, and Microsoft Ads…
Analyze Fireflies meetings and Circle community posts, for one person, a team, or everything, then produce a weekly or daily Customer Intelligence Report as a BenAI-branded HTML dashboard in the instant-ui design language. Surfaces recurring themes across a trailing window, pain
$ npx -y skills add naveedharri/benai-skills --skill customer-intelligence-report --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/customer-intelligence-reportContext preview
The summary Claude sees to decide when to auto-load this skill.
Analyze Fireflies meetings and Circle community posts, for one person, a team, or everything, then produce a weekly or daily Customer Intelligence Report as a BenAI-branded HTML dashboard in the instant-ui design language. Surfaces recurring themes across a trailing window, pain
name: customer-intelligence-report description: "Analyze Fireflies meetings and Circle community posts, for one person, a team, or everything, then produce a weekly or daily Customer Intelligence Report as a BenAI-branded HTML dashboard in the instant-ui design language. Surfaces recurring themes across a trailing window, pain points, technical patterns, and stuck members needing follow-up, each backed by verbatim quotes. Blends both connectors, degrades gracefully if one is down, stores no secrets. Fully data-driven and customizable: sources, sections, scope, cadence, and recurring window are all configurable. Use when the user says \"customer intelligence report\", \"community pulse\", \"analyze our meetings\", \"weekly community report\", \"what themes came up in calls\", \"who is stuck\", \"Fireflies report\", \"Circle report\", or \"meeting intelligence\"." disable-model-invocation: true
Turn a period of Fireflies meetings and Circle community posts into a recurring intelligence report: what themes keep coming up, where people are in pain, which technical patterns recur, and who is stuck. The deliverable is a BenAI-branded instant-ui HTML dashboard rendered from a data file, so it is customizable and never static (see `references/report-config.md`).
**Secrets:** the Circle and Fireflies connectors authenticate through the MCP connection. Never put an API key, token, space id, or email into the skill as a hardcoded secret, all targeting values come from the run config. If you add publishing, use an env-var token and a gitignored deploy dir (see `report-config.md`, Secrets and safety).
Track progress:
Task Progress: - [ ] 1. Scope: confirm scope, cadence, sections (checkpoint) - [ ] 2. Collect: pull the period from Fireflies and Circle - [ ] 3. Analyze: cluster transcripts into themes - [ ] 4. Assemble: build the WEEKS + LEXICON data - [ ] 5. Build: render the instant-ui dashboard
Confirm the run with the user. Present the options and wait for a pick, do not assume. Read `references/report-config.md` for the choices. Confirm: sources (Fireflies, Circle, or both, plus their targeting: organizer emails and Circle space ids), scope (all / person / team), cadence (weekly / daily / custom), the trailing recurring window, and which sections to include. Offer the default section catalog plus 3 to 5 optional sections so they can shape it. Targeting values are config, never hardcoded secrets.
Pull the period from the configured sources. Both are optional and degrade gracefully: if one errors, continue with the other and note the outage.
List meetings and posts first and report the counts to the user, then fan the transcript and post reading out to subagents that return small structured notes.
Cluster the notes into themes. Read `references/analysis-framework.md`. Assign each theme a section, canonicalize its id against the LEXICON, compute counts, streak, and resolved, and detect stuck members. Every theme needs a real verbatim quote or it is cut.
Shape the analysis into the data contract. Read `references/data-schema.md`. Produce valid JSON for `WEEKS` (append this period to prior periods if they exist) and update `LEXICON` with any new phrasings. Present a short chat summary of the top themes and stuck members.
Render the report in the BenAI instant-ui design language. Load the embedded guides first: read `references/instant-ui/design-tokens.md` (the `:root` token block and typography rules), `references/instant-ui/page-shell.md` (gradient stripe, dark BenAI header, smiley SVG, breakpoints), `references/instant-ui/components.md` (copy component CSS/HTML verbatim if a section needs a block beyond the template), and `references/instant-ui/build-rules.md` (hard constraints and the after-build checklist). Then read `references/dashboard-template.html`, which is already built in this language: set the config constants and `SECTIONS`, paste in the `WEEKS` and `LEXICON` JSON, and write the filled file to `customer-intelligence-report-YYYY-MM-DD.html` in the working directory. The dashboard must be self-contained and offline, and must keep the instant-ui look: cream `#fffef8` background, 3px solid `#111` borders, hard zero-blur shadows, square corners, weight-900 headings, SF Mono for every number and label, the 4px gradient stripe, and the BenAI smiley in header and footer. Run the after-build checklist before confirming. No em dashes in the rendered copy. Never invent a number or quote, every figure traces to a real Fireflies or Circle result.
This skill is never finished. Improve it as you use it.
reference file (or this SKILL.md) so the correction sticks.
emerging entries, so recurrence tracking improves over time.
as a model for future runs.
no longer changes behavior.
| Step | Reference | |------|-----------| | 1. Sources, scope, cadence, sections | `references/report-config.md` | | 2. Collect meetings (Fireflies) | `references/fireflies-guide.md` | | 2. Collect posts (Circle) | `references/circle-guide.md` | | 3. Analyze into themes | `references/analysis-framework.md` | | 4. Data contract | `references/data-schema.md` | | 5. Dashboard | `references/dashboard-template.html` + `references/instant-ui/` guides |
Expert automation skills for Claude Code, organized by department.
Repo: naveedharri/benai-skills
Full multi-platform paid advertising audit with parallel subagent delegation. Analyzes Google Ads, Meta Ads, LinkedIn Ads, TikTok Ads, and Microsoft Ads…
Budget allocation and bidding strategy review across all ad platforms. Evaluates spend distribution, bidding strategy appropriateness, scaling readiness, and…
Competitor ad intelligence analysis across Google, Meta, LinkedIn, TikTok, and Microsoft. Analyzes competitor ad copy, creative strategy, keyword targeting,…
Cross-platform creative quality audit covering ad copy, video, image, and format diversity across all platforms. Detects creative fatigue, evaluates…
Google Ads deep analysis covering Search, Performance Max, Display, YouTube, and Demand Gen campaigns. Evaluates 74 checks across conversion tracking, wasted…
Landing page quality assessment for paid advertising campaigns. Evaluates message match, page speed, mobile experience, trust signals, form optimization, and…