agentleFS
moronireismoronireis/agent-system-partners/.claude/skills/pipeline-copy/SKILL.md

pipeline-copy

Skill0 starsChanged 2 months agoMITView on GitHub

The direct-response copy pipeline — mkt-strategist (brief) to mkt-copywriter to the mkt-editor gate to sign-off. Use to run or explain the end-to-end chain for sales pages, email sequences, ad copy, or VSL scripts.

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:

    No unresolved veto conditions remain.

    line 45
  2. Very specific:

    When this pipeline is one task among several in a plan, the task contract's Gate field should read mkt-editor, and the Success criteria …

    line 59

≈ “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

Pipeline: Copy

Chain for direct-response copy (sales pages, emails, ad copy, VSL scripts) from strategic brief to sign-off.

Copy in this pipeline is judged on persuasion and conversion, not reach — that's the dividing line against pipeline-content below.

Chain

mkt-strategist (brief)
  → mkt-copywriter (raw persuasive copy)
    → mkt-editor (humanize + quality gate)
      → PASS → mkt-strategist (strategic sign-off) → done
      → BLOCK → named rollback agent (usually mkt-copywriter) → re-enter mkt-editor
      → EXIT (2 loops exhausted or critical failure) → escalate to director/user

Steps

  1. mkt-strategist loads the project pack and any existing research (brain/ops/handoffs/research-*.md), then writes a brief to brain/ops/handoffs/brief-<slug>.md: ICP/segment, positioning angle, offer reference, format requested, key proof points, constraints.
  2. mkt-copywriter reads the brief and the project pack's voice.md, loads the copywriting-frameworks skill, and produces raw copy. Persuasion is maximized here — naturalness is not this agent's job.
  3. mkt-editor loads the humanization skill and the project pack's voice.md, runs the humanization pass, then scores the piece against its veto conditions and issues a verdict to brain/ops/reports/copy-<slug>.md.
  4. On PASS, mkt-strategist gives final strategic sign-off; the piece is ready for its destination (publication is a human action, or the piece is handed to dev-builder / mkt-ads depending on format).
  5. On BLOCK, the verdict names the rollback agent and the exact fix needed; that agent redoes only its own step, then the piece re-enters mkt-editor.
  6. On EXIT, or after 2 revision loops without a PASS, escalate to the user with the blocker stated plainly — never attempt a 3rd silent loop.

When to Use

  • Any deliverable whose job is to directly persuade toward a single action: sales/landing pages, email sequences, ad primary text, VSL scripts, application-page copy.

When Not to Use

  • Social content whose primary job is reach/engagement rather than direct conversion — use pipeline-content instead (it still gates through mkt-editor, but originates from mkt-content, not mkt-copywriter).
  • Anything that skips the strategic brief — mkt-copywriter never invents strategy; if there's no brief, the pipeline starts at step 1, not step 2.

Done When

  • The mkt-editor verdict is PASS and logged at brain/ops/reports/copy-<slug>.md.
  • mkt-strategist sign-off is recorded.
  • No unresolved veto conditions remain.

Parallelization

Independent deliverables (three unrelated emails, ad copy for two different angles) can each run their own brief → copywriter → editor chain in parallel — they don't share state. A single multi-part deliverable (one sales page with several sections) stays sequential within itself; don't split one page across parallel writers and reassemble, voice and argument continuity break.

Common Deviations

  • Brief is incomplete (no ICP, no offer reference, no format specified) — mkt-copywriter reports back to mkt-strategist rather than guessing at strategy. This is a contract violation, not a normal input, and it is reported, not silently absorbed.
  • Editor traces a BLOCK back to the brief itself, not the execution (e.g., the requested angle contradicts the project pack's positioning) — escalate directly to mkt-strategist instead of cycling mkt-copywriter against a broken brief.
  • A revision loop fixes one veto but introduces another — this still counts against the 2-loop cap; it is not a fresh count.

Notes for the Director

When this pipeline is one task among several in a plan, the task contract's Gate field should read mkt-editor, and the Success criteria should require the specific verdict file path — not just "copy written."

Independent copy contracts (different slugs, different offers) are safe to run in parallel per docs/task-contract.md's "one responsibility per contract" rule. Never collapse two unrelated deliverables into a single contract just to save a delegation — it makes the eventual BLOCK harder to trace back to a single owner.

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.