adopt
Use when an existing project that was not built with X2 must be migrated to run under it, with documentation sprawl to dissolve, stories to harvest at epic…
Use when the builder believes the slice is green, to produce the slice-exit report that is the definition of green. Triggers before any green declaration and at slice close. Green is declared by the report and ratified only after intake audits it. Never declare a slice green in
$ npx -y skills add teklabsdigital/x2-method --skill slice-exit --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/slice-exitContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when the builder believes the slice is green, to produce the slice-exit report that is the definition of green. Triggers before any green declaration and at slice close. Green is declared by the report and ratified only after intake audits it. Never declare a slice green in
name: slice-exit description: Use when the builder believes the slice is green, to produce the slice-exit report that is the definition of green. Triggers before any green declaration and at slice close. Green is declared by the report and ratified only after intake audits it. Never declare a slice green in prose, and never skip a section because it is inconvenient.
The definition of green. The pilot declared "all green" twice before it was true, and the true state had to be extracted by interrogation; every "confirm X" the human issued was the same turn shape, repeated because the method had no artifact in which the builder declares, with evidence, what green means. This report is that artifact, and it has teeth: a slice without a complete report is not green (MET-06).
1. **Tiers, run this session, with counts.** Every tier against the final tree: build, unit, architecture, integration, e2e, client verification. A stale last-known-good is marked as stale, never presented as run. 2. **Composed-app smoke result.** The actual entrypoint, at least one real request per primary flow. The pieces passing their own tests is not the product running (INV-07). 3. **Service-method coverage table.** Method, UI surface, harness scenario; gaps named. A method without UI in the locked portion is reported as the sanctioned UI-unsurfaced state with its e2e scenario and boundary note (CONF-03), so the skew is visible and deliberate. 4. **Unsurfaced decisions.** Must be none. Anything the build decided that the human never saw is listed here now, not discovered later. 5. **Pending rulings owed to the human.** 6. **Prototype-lock status.** The lock on record, provenance stamped, any owed fidelity items named (UI-4). 7. **The slice turn count, from the ledger.** The ledger file governs; a count remembered by the builder does not (MET-05).
Plus two standing checks: any phase that started without its gate on record (stories, prototype approval, the lock) is a methodology violation and is surfaced here (MET-07); and the report names every other slice built past its own report, so exit debt is visible, never silent. The third project wrote one report two slices late and had four owed by extraction time; each of those slices was invisible to the metric until its report existed.
Scripted where mechanical (tier counts, the coverage diff), authored where judgment is needed (deviations, pending rulings).
Intake audits the report before ratification. The builder does not grade its own green; "sound" is intake's word to give. The human's turn becomes reading one document, not extracting it. What the human reads is put in plain language, their own words; method vocabulary rides alongside, never instead, because a ratification given on a summary the owner cannot parse is no ratification (P3).
Where the report lives: the pilot used the repo's working docs; the alternative is the extraction record. Carried on the record; the pilot's `docs/work/` is the working default until ruled.
the report is produced again from the top.
An agentic software engineering method: AI agents do the building, the build enforces the rules that must always hold, a person decides only at the gates, and the method measures its own cost in human turns.
Repo: teklabsdigital/x2-method
Use when an existing project that was not built with X2 must be migrated to run under it, with documentation sprawl to dissolve, stories to harvest at epic…
Use after the repo is seeded to run the D-000 five-lens decomposition, record persistent schema decisions, define slice one, and start the deltas file. Also…
Use only after the lock (gate 2b) is on record for the slice portion, to derive the fidelity ledger and acceptance scenarios from the locked prototype.…
Use when D-000 exists and the whole-product prototype does not, to run Claude Design in its own context and produce the prototype and design system. Also use…
Use when an X2 project completes (or reaches a major milestone) to run the extraction loop, the quality feedback that folds the project's evidence back into…
Use when someone asks what an X2 skill does, which skill to run next, how the gates work, or where an artifact lives. A quick in-flight reference; for the…