tactical-model
Skill1 starsChanged 3 months agoMITView on GitHub
Defines the content and structure of a Slice's Tactical Model — aggregates, invariants, domain events, commands, and policies. Reference when assembling or reviewing slices/<id>/tactical-model.json.
In this repository17 other files it gives its agents.
- acceptance-criteriaSkill
- adjudication-recordSkill
- adrSkill
- behavior-inventorySkill
- characterization-suiteSkill
- context-modelSkill
- contract-testsSkill
- cutover-planSkill
- design-docSkill
- divergence-registerSkill
- interface-inventorySkill
- migration-waypointsSkill
- observability-specSkill
- retirement-registerSkill
- roadmapSkill
- slice-outlineSkill
- variation-policySkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 31
If an invariant is suspected but its exact statement isn't settled, do not add a placeholder invariant just to satisfy the "≥1 invariant" …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
Tactical Model
Defines what a Tactical Model contains. Does not decide the boundaries — those were fixed at Stage 2; this is the per-Slice detailed design within an already-approved Context, just-in-time (07-stages.md Stage 5).
Schema
schemas/tactical-model.schema.json. A single JSON object per Slice — slices/<id>/tactical-model.json is not an array.
Required fields
slice, aggregates (≥1, each with name, root, invariants).
Content obligations — Vernon's rules, applied (knowledge/ddd-reference.md §2.1)
- Every aggregate declares at least one invariant. An aggregate with no invariant is a data bag, not a domain model — the Stage 5 gate rejects this (
trace-lint.py --stage 5). The test: can you state the rule that must hold at the end of every transaction? If not, it isn't an aggregate. - Reference other aggregates by identity only — never embed one aggregate inside another's structure.
- No invariant should require a synchronous call to a sibling aggregate. This is the specific judgment criterion the Stage 5 architecture-review approval must cover explicitly.
- Domain events are named in the past tense (
OrderPlaced, notOrderTableUpdated); commands are named in the imperative (PlaceOrder). A domain expert should recognize the event name.
Behavior-ID citation
Every invariant's behaviors[] must be non-empty — this is both "every aggregate declares ≥1 invariant" and "every design clause cites a Behavior ID" from the Stage 5 gate, in one field. A Behavior assigned to this Slice (per its slice.json.behaviors[]) that is cited by no invariant here is an orphan Behavior in the Slice — also a gate failure.
Unknowns
If an invariant is suspected but its exact statement isn't settled, do not add a placeholder invariant just to satisfy the "≥1 invariant" check — an invariant that doesn't actually hold is worse than an aggregate flagged as needing more design work.
Minimal worked example
{
"slice": "SLC-0004",
"aggregates": [
{
"name": "Order", "root": "Order",
"entities": ["OrderLine"], "value_objects": ["Quantity", "Money"],
"invariants": [
{ "statement": "An Order must contain at least one OrderLine with quantity > 0", "behaviors": ["BHV-0042"] }
]
}
],
"domain_events": [ { "name": "OrderPlaced", "aggregate": "Order" } ],
"commands": [ { "name": "PlaceOrder", "aggregate": "Order" } ],
"policies": [ "When OrderPlaced, notify billing-svc" ]
}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
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.

