| name | gtm-repo |
| description | How this GTM repository works. Load when orienting in this repo, deciding where a change belongs, or explaining the layer conventions to someone. Covers the layers, the dependency direction, the PR workflow, the confidence ladder, and where each kind of change goes. |
Operating this GTM repo
This repo follows the Manifest framework (Manifest, by Cargo). Strict dependency
direction: infra reads context, skills read everything, nothing depends on
outputs.
Where does a change go?
| You want to... | Touch |
|---|
| Set the goal, the forecast, or what we are not doing | plan/ |
| Start a bounded effort with a deadline and success criteria | initiatives/<name>.md |
| Plan the week, log the day, track a dropped ball | cadence/ |
| Update ICP, personas, positioning, objections, proof | context/<domain>/*.md |
| Change what runs in production (models, tools, plays, agents) | infra/**/*.ts |
| Add a repeatable procedure for agents working in this repo | .claude/skills/<name>/SKILL.md |
| Run a one-off operation against the workspace | scripts/*.ts |
| Guard an agent prompt against regressions | evals/<agent>/ |
| Archive the result of a run, campaign, or research task | outputs/YYYY-MM-DD-<slug>/ |
| Keep a half-baked experiment to yourself | scratch/ (gitignored) |
Rules that are easy to miss
- The confidence ladder. Files in
proof/, signal/, and insight/ carry
confidence: hypothesis | validated | proven. Outreach motions may cite only
validated or proven. The linter enforces it. Use the capture-feedback
skill to promote a claim, and only on a second independent occurrence.
- Outreach motions declare a hypothesis before any copy exists: segment,
angle, trigger, expected_reply_rate. Enrichment, scoring, routing, and signal
motions are not campaigns and are exempt: they declare intent through the
outcome they serve in
plan/outcomes.md. Set kind: on every motion.
- Motion vs play. A motion is the GTM move; its narrative lives in
context/motion/. A play is a Cargo resource (definePlay: watches a
data model, runs per row). A motion is implemented as a Cargo play, tool, or
agent. Do not confuse them.
- Every outcome in
plan/outcomes.md names what it becomes in infra/. An
outcome with no Becomes: is a wish.
outputs/ entries carry an outcome: field (meetings, replies, pipeline
attributed), so motions are ranked by results rather than opinions.- Absolute dates everywhere.
2026-07-14, never "last Tuesday".
- Outputs are append-only. Correct with a dated addendum. Major context
rewrites archive the prior version instead of overwriting it.
infra/cargo.state.json is machine-owned. Never edit it or resolve its
conflicts by hand; re-run the deploy instead.- Every context domain has a
_template.md. Copy it; do not invent frontmatter
fields.
- Deploys happen from CI on merge to main, never locally.
- Copy style everywhere: no em dashes.
Checks before opening a PR
npm run lint (context, outputs, and copy style) if you touched context/,
outputs/, or any prose.npm run typecheck if you touched infra/ or scripts/.npm run eval if you touched a system prompt (needs OPENAI_API_KEY).
