The 22 skills · Platform skill · Everything only
AGENT SKILL
delivery-and-alerts: what it makes your agent do, and what it refuses.
delivery-and-alerts covers how reporting is actually consumed: the Monday email, the channel post, the alert that fires when a line is crossed. Most of a dashboard's audience never opens the dashboard, so push is the product, and the same discipline has to ride along with it.
The job delivery-and-alerts does
The chain runs in one order: refresh, prove, then send. A scheduled refresh rebuilds from the landing zone and re-runs the gates, so nothing is delivered that the build did not certify. Subscriptions carry the figure, the as of date and the check state into email and chat, rather than a screenshot that ages silently. Alerts evaluate at refresh, deterministically, from processed data, with no polling loop and no model deciding what looks anomalous. Thresholds live in versioned configuration with an owner recorded against each line, and the message names both the rule and its owner so a firing alert has somewhere to go.
What delivery-and-alerts refuses
A skill is worth what it stops as much as what it produces. These are the refusals written into delivery-and-alerts, taken from the rules and limits in its own file.
- ✓Sending what the build did not certify. Delivery runs after the gates, never around them.
- ✓An alert owned by nobody. Every threshold is a recorded decision with an owner, so it does not die when one person leaves.
- ✓A black box anomaly ping. Conditions are explicit and versioned, because an alert nobody can explain is an alert nobody trusts.
- ✓A rule left unreviewed. A line that never fires and a line that fires daily are both asking for a decision.
9 lines from the delivery-and-alerts skill file, verbatim
From skills/delivery-and-alerts/SKILL.md, unedited. Kymira skills are plain text you can
read in full before you install them, which is the point of shipping a method rather than a binary.
- **Alerts evaluate at refresh, deterministically, from processed data.** No polling loop, no second query path, no model deciding what's anomalous. The refresh that proved the numbers also checks the lines. - **Thresholds are config, not code** (doctrine #8): a versioned `alerts.json` — metric key, comparator, line, audience, channel. Every threshold is a decision someone recorded; the alert message names the rule AND its owner. - **Team-owned, not personal.** Alerts route to channels/groups from the config file. No alert dies with an inbox.
The failure modes delivery-and-alerts addresses
No entry in the catalogue of 51 observed failure modes names delivery-and-alerts as the thing that addresses it. The catalogue records failures we have watched happen, so a skill appears there once a failure it covers has actually been observed and written up. Read the catalogue.
The tier delivery-and-alerts ships in
| Tier | Price | What the tier holds | delivery-and-alerts |
|---|---|---|---|
| Free | $0 | honest-dataviz and the eleven doctrine principles, published in full | Not in this tier |
| Core Pack | $199 | the thirteen craft skills, one-time purchase, lifetime updates | Not in this tier |
| Everything | $349 | all 22 skills, both libraries, the template bank, one-time purchase, lifetime updates | Included |
Both paid tiers are a one-time purchase with lifetime updates, and all sales are final: the files are plain text you keep the moment they arrive, so judge the quality on the free skill and the published doctrine first. The pricing section states exactly what each tier contains.
Installing delivery-and-alerts
A skill is a folder with a SKILL.md file in it. Installing one is copying that folder where
your agent looks for skills, with no build step and no dependencies.
mkdir -p ~/.claude/skills cp -R Kymira/skills/delivery-and-alerts ~/.claude/skills/
Or open the unzipped pack in Claude Code and say install Kymira, which installs every skill in your tier, or double-click the installer for your system. Cursor reads the same skill as a rule file: copy delivery-and-alerts.mdc into .cursor/rules/. Skills are read at startup, so start a new session and ask your agent which skills it has available.
Not for
Delivery proves sending, not reading, so pair it with the owner's usage view before concluding a report matters. Thresholds encode judgment: this skill makes them explicit, versioned and owned, and cannot make them right. It is a platform skill, so it ships in Everything only.
Last updated 2026-08-14 · Dylan, founder · one of 22 packaged skills, each one a method written down, not a product tour.