Goal coverage

The outcome

Nobody has to know the catalog to know what is missing. Once a week one agent compares the goal with what the workspace runs and says, in the workspace's own numbers, what it noticed: "you have 4,200 accounts and no score on any of them". Each observation comes with the pipeline that fixes it. Two things land:

1. A coverage file, cadence/coverage/<YYYY-Www>.md, in one pull request titled [cadence] coverage YYYY-Www. It holds the goal verbatim, what runs today, and at most five suggestions, each with a status: open line. To decline one, the reviewer changes that line to status: declined: <reason> before merging. That edit is the agent's memory: a declined suggestion stays out for four weeks, and one left open for three weeks is dropped as noise.
2. A Slack post of the same suggestions, linking the pull request, for the reader who never opens GitHub.

It never installs what it suggests. The install line is in the file, and running it is a human's decision.

What it is not. It does not grade the work that already exists: whether the bets you have are on track is weekly-planning, and whether a deployed play is failing is workspace-health. An installed pipeline counts as covered, however badly it is going.

Three properties make it safe to run unattended:

• The menu is fetched, never remembered. The agent reads https://raw.githubusercontent.com/getcargohq/gtm-skills/main/catalog.json every run and suggests only names it finds there. If the fetch fails it suggests nothing.
• Installed is decided by three signals, not a guess: infra/<name>/ in the checkout, a workspace folder named after the pipeline (every pipeline files its resources under one), or a deployed play or agent slug matching the name.
• Read-only, then a pull request. Every workspace read is a named cargo-ai list or an aggregate query. Its write paths are the pull request and one locked Slack channel.

Example

Every week, look at our goal and what we have running, and tell us which pipelines we are missing.

Illustrative output, fictional records:

Goal coverage, 2026-W44: "Reach $8M ARR by 2026-12-31, up from $5.2M on 2026-06-30."

1. Outbound to mid-market is a third of the forecast and nothing tells you
   when to reach out: 1,840 tier A accounts, none watched for a new hire. Flag
   the week a tier A account hires the role you sell to (`new-hire-detection`).
2. Expansion is $1.1M of the plan and no account is watched for it: 312
   customers, no usage or seat signal on any. Flag the customers ready to grow
   (`expansion-signals`). Needs the HubSpot connector first.
3. Still missing, first suggested 2026-W42: every demo request answered on the
   page (`inbound-qualification`).

Decline or adopt in the pull request: github.example/acme/gtm/pull/212

Three suggestions where five were allowed, each opening on a number from the workspace, and the scoring model declined in W43 is not among them.

Put it in your project

This folder is a worked example: real project resources written for some other company. The job is to end up with the code your company would have written, in your project, and an agent does the adapting. If the cargo-project skill is in your session it carries the long form of this; if not, this is enough.

1. Install it: the CLI does the copy. From inside the project (the folder holding cargo.json), cargo-ai project add cookbook/goal-coverage writes this example to infra/goal-coverage/ and this procedure to .agents/skills/goal-coverage/. No project yet? cargo-ai project init <dir> --cookbook goal-coverage && cd <dir> && npm install does both. If you are reading this from the project's .claude/skills/, the install already happened, so start at step 2. On a CLI too old to have add, copy this folder in as a sibling of what is there by hand; everything below is unchanged.
2. Reconcile it with what is already declared. If the project already has an Anthropic, GitHub or Slack connector, rewire the imports to the existing one and drop the copy; two resources with one slug is a collision at deploy.
3. Adapt. Work the sections below in order: What should not change is what you argue back about; What you can change is what you offer unprompted; What you will be asked is the floor, and you derive before you ask. Record what you changed and why under a ## Decisions section in your copy of this file.
4. Check, then plan. node --import tsx evals/contract.mjs from this skill's folder, then npm run check && cargo-ai project plan from the project root, show the diff, and deploy only on an explicit yes: cargo-ai project deploy (or merge, where a merge deploys). Never cargo-ai project init --force into a non-empty directory.
5. Verify. Run the agent once without waiting for Tuesday, then walk Done when line by line and report each with evidence.

What you will be asked

Derive before you ask. Only the rows marked asked genuinely live in the operator's head.

What you can change

What should not change

• The sandbox runs as a viewer service user. (infra/users/index.ts) With no runAs the cargo-ai session is missing and every workspace read fails; with a broader role a prompt that is ignored can deploy or edit a resource.
• It suggests only what the fetched catalog lists. (goal-coverage.prompt.ts §3) A pipeline named from memory may have been renamed or removed, and the install line then fails on the one command the reader was persuaded to run.
• It never installs, deploys or enables anything. (goal-coverage.prompt.ts §2, Never) The suggestion is a sentence; the install is a decision with a cost the agent does not see (a connector to authorize, credits on every run).
• Every observation carries a number it read. (goal-coverage.prompt.ts §5) "You may want scoring" is a pitch; "4,200 accounts, none scored" is a finding. A number it did not read is worse than none, because the reader repeats it.
• Installed is covered, however badly it is going. (goal-coverage.prompt.ts §1) Grading existing work is weekly-planning's and workspace-health's job. Do it here too and the same failing play is reported three times by three agents.
• The decline memory is the coverage file. (goal-coverage.prompt.ts §4) Drop the pull request and a declined suggestion returns every Tuesday until someone mutes the post.

Done when

• node --import tsx evals/contract.mjs passes: a claudeCode harness bound to an Anthropic connector and a model, runAs a viewer service user, no capability, exactly one connector action (slack.postMessage with the channel locked), and a COVERAGE_GOAL env entry
• cargo-ai project plan reports the agent, its folder, and the three connectors
• the first run opened one unmerged pull request titled [cadence] coverage <YYYY-Www> whose diff is exactly cadence/coverage/<YYYY-Www>.md, with the goal quoted verbatim and at most five suggestions, each ending status: open
• no suggestion names a pipeline that is installed by any of the three signals, or one that is not in the catalog
• every observation's number appears in a read listed under ## Checked
• the Slack post went to the locked channel, once, and links the pull request
• a second run the same week opened nothing and posted nothing
• after a suggestion was merged as status: declined: <reason>, the next week's run did not suggest it

What it costs

The recurring cost is the harness run once a week, billed as LLM tokens through the bound Anthropic connector. It scales with what the agent reads: the plan file, a few list commands, one or two aggregate queries, the catalog, and up to eight earlier coverage files. The workspace reads are API calls, not connector actions, so they bill nothing, and the run never executes an action that spends. The Slack post is one connector action a week.

Composes into

Whatever it suggests. Each suggestion's install line is cargo-ai project add cookbook/<name>, and the next run counts that pipeline as covered. Pairs with weekly-planning (are the bets we have on track) and workspace-health (is what we deployed broken).