agentleFS
styfinitystyfinity/claude-marketing-department/skills/research-to-brief/SKILL.md

research-to-brief

Skill6 starsChanged 4 months agoMITView on GitHub

Folds any pile of research — pasted notes, files, transcripts, reports — into the right sections of marketing-brief.md, graded T1 to T3 for confidence, with contradictions surfaced instead of averaged away. Use whenever research exists anywhere other than the brief.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write the grade inline with the finding.

    line 22
  2. Very specific:

    Merge into marketing-brief.md: Update sections in place, date-stamp the pass, and keep customer quotes verbatim in the language bank.

    line 24

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Research to Brief — one home for what you know

Research that lives in scattered documents might as well not exist: the next campaign gets built from memory and vibes regardless. The brief is the department's single memory, and this skill is its intake valve — everything decision-relevant gets extracted, graded, routed to its section, and reconciled with what is already there.

Inputs

  • The research, in any form — pasted text, file paths, folders, notes: $ARGUMENTS
  • The whole current marketing-brief.md, read end to end first, so new material lands as an update rather than a duplicate. If no brief exists, run /marketing:onboard to create the skeleton, then return here.

Do this

  1. List the sources with dates. Recency matters at grading time, and every finding must trace back to one of them.
  2. Extract every decision-relevant finding — a claim, a number, a verbatim quote, a pattern — and route each to its brief section: the business, the market, the customer (ICP, avatar, pains, desires, objections, language bank), the competitors, positioning & message, the offer, channels & funnel, voice, numbers that matter. A finding can feed two sections; cross-reference rather than duplicate.
  3. Grade each finding on the brief's own scale:
    • T1 sourced — a customer said or did it, or a measured result.
    • T2 reasoned estimate — credible secondhand: review mining, reports, competitor observation.
    • T3 guess — inference, or a single anecdote doing a lot of work. Write the grade inline with the finding.
  4. Surface contradictions. Where new findings disagree with the brief or with each other, do not overwrite and do not average — log both versions in a "contradictions to resolve" block, each with the one input that would settle it.
  5. Merge into marketing-brief.md. Update sections in place, date-stamp the pass, and keep customer quotes verbatim in the language bank. An existing entry is replaced only when the new evidence grade is equal or better — a T3 never overwrites a T1.
  6. Report the delta: what was added to which section, what got upgraded or downgraded, the contradictions raised, and the single research task that would most improve the brief's weakest section.

Output

An updated marketing-brief.md, plus a change log in chat: findings added per section with grades, contradictions to resolve, and the one next research task. The brief's T3-marked sections remain the visible research backlog.

Rules

  • Verbatim beats paraphrase. Customer language enters the language bank exactly as spoken — the copy skills depend on it.
  • Grade honestly. A brief full of T3 guesses dressed as T1 facts briefs the whole department into fiction.
  • Contradictions are findings. Averaging two conflicting numbers destroys both.
  • Nothing enters without a source. "Everyone knows" is not a source.
  • Date-stamp every pass. An undated finding cannot be superseded cleanly, and stale certainty is worse than fresh doubt.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.