@next-friday/create-app 9.2.1 → 9.3.0

package/templates/base/.agents/skill-evals/sdlc/fix-reproduces-finds-the-cause-and-waits-for-the-engineer/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 30
-timeout_seconds: 900
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc fix a bug: parseVersion("1.2") in src/version.ts returns "1.NaN.0" where it should return "1.2.0"

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/case.yaml

@@ -1,11 +0,0 @@
-schema_version: "1.1"
-name: interviews-before-writing-intent
-expected_outcome: |
-  The agent restates the goal, picks the plan playbook, and asks numbered questions that each carry a recommended answer. It writes no intent yet.
-timeout_seconds: 600
-checks:
-  - reply_matches: '^\s*(?:\*\*)?1[.)]'
-  - reply_matches: '^\s*(?:\*\*)?2[.)]'
-  - no_new_commits
-  - no_new_branch
-  - reply_lacks: '(?:^|[\s`("''])/(?:home|tmp|root|Users|var|private)/'

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/01-the-first-message-restates-the.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The first message restates the goal, the problem, and the smallest change expected to meet it

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/02-the-reply-asks-numbered-questions.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply asks numbered questions; each carries a concrete recommended answer that "yes" could accept, or says that it has none to recommend

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/03-no-intentmd-is-written-before.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No `intent.md` is written before the originator answers

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/04-reply-lists-the-playbook-steps.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply's Steps block lists the plan playbook's seven steps, each copied word for word, and marks each one done, open, or `skip:` with a reason

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/05-no-commit-is-made-nothing.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No commit is made, nothing is pushed, and no pull request is opened

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/07-every-path-in-the-reply.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-Every path in the reply starts from the project root; none starts from the machine's root

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/graders/08-no-question-in-the-round.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No question in the round depends on the answer to another question in the same round

package/templates/base/.agents/skill-evals/sdlc/interviews-before-writing-intent/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 20
-timeout_seconds: 600
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc customers complain that the login page is very slow, and I want it fixed

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/case.yaml

@@ -1,14 +0,0 @@
-schema_version: "1.1"
-name: maintain-writes-intent-from-a-finding
-expected_outcome: |
-  A job reports that GET /health answers 500 and no person is available. The agent follows the maintain playbook: it diagnoses without changing a source file, writes intent/<change>/intent.md from the template with the error text and where it was seen under Problem, puts an unverified cause under Open questions, pushes a branch, and names the service owner's gate. It asks the person no question.
-timeout_seconds: 900
-checks:
-  - file_matches:
-      path: intent/*/intent.md
-      regex: 'Cannot read properties of undefined \(reading .status.\)'
-  - file_matches:
-      path: intent/*/intent.md
-      regex: '(?=[\s\S]*^## Problem$)(?=[\s\S]*^## Proposed outcome$)(?=[\s\S]*^## Affected users and systems$)(?=[\s\S]*^## Constraints$)(?=[\s\S]*^## Open questions$)'
-  - commit_touches_only: "intent/**"
-  - branch_matches: "^(?!main$).+"

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/fixture.md

@@ -1 +0,0 @@
-A generated NestJS project with a local `origin`; `src/health/health.controller.ts` exists. The finding is the job's claim, and the project holds no seeded fault. No person is available to answer questions.

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/graders/01-the-intent-has-the-five-headings.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-`intent/<change>/intent.md` exists with exactly the template's five headings: Problem, Proposed outcome, Affected users and systems, Constraints, and Open questions

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/graders/02-problem-holds-the-error-and-where.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-`intent/<change>/intent.md` holds the error text "Cannot read properties of undefined (reading 'status')" and where it was seen, `GET /health` and `src/health/health.controller.ts`, in its Problem section

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/graders/03-unverified-cause-is-an-open-question.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-A cause that the session did not verify is under Open questions, not stated as fact in Problem or Proposed outcome

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/graders/04-no-source-file-is-changed.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No file outside `intent/` is changed

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/graders/05-a-branch-is-pushed-and-the-gate-is-named.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-A new branch holding `intent.md` is pushed, and the reply names the service owner as the gate that triages the finding

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/graders/06-the-person-is-asked-no-question.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The session asks the person no question

package/templates/base/.agents/skill-evals/sdlc/maintain-writes-intent-from-a-finding/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 30
-timeout_seconds: 900
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc an automated check found that GET /health answers 500 every time since the latest commit. log: "TypeError: Cannot read properties of undefined (reading 'status')" at src/health/health.controller.ts. Nobody is available to ask.

package/templates/base/.agents/skill-evals/sdlc/reported-problem-starts-the-plan/case.yaml

@@ -1,11 +0,0 @@
-schema_version: "1.1"
-name: reported-problem-starts-the-plan
-expected_outcome: |
-  The message reports a problem through /sdlc without asking a question. The agent follows the plan playbook: it lists the steps and asks its first round. It does not tell the person to send /sdlc again.
-timeout_seconds: 600
-checks:
-  - reply_matches: '^\s*(?:\*\*)?1[.)]'
-  - reply_lacks: '(?<!skills)/sdlc(?![\w/-])'
-  - no_file_changes
-  - no_new_commits
-  - reply_lacks: '(?:^|[\s`("''])/(?:home|tmp|root|Users|var|private)/'

package/templates/base/.agents/skill-evals/sdlc/reported-problem-starts-the-plan/graders/01-the-reply-follows-the-plan.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply follows the plan playbook: it has a Steps block and asks at least one numbered question

package/templates/base/.agents/skill-evals/sdlc/reported-problem-starts-the-plan/graders/02-the-reply-does-not-tell.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply does not tell the person to send `/sdlc` again to start the work

package/templates/base/.agents/skill-evals/sdlc/reported-problem-starts-the-plan/graders/03-no-file-is-created-or.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No file is created or changed, no commit is made, and nothing is pushed

package/templates/base/.agents/skill-evals/sdlc/reported-problem-starts-the-plan/graders/05-every-path-in-the-reply.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-Every path in the reply starts from the project root; none starts from the machine's root

package/templates/base/.agents/skill-evals/sdlc/reported-problem-starts-the-plan/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 20
-timeout_seconds: 600
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc the login page is very slow and customers complain

package/templates/base/.agents/skill-evals/sdlc/request-that-fits-no-row/case.yaml

@@ -1,10 +0,0 @@
-schema_version: "1.1"
-name: request-that-fits-no-row
-expected_outcome: |
-  The request matches no row of the /sdlc table. The agent says which row is closest and why it does not fit, answers in a few lines, and stops. It writes no file and commits nothing.
-timeout_seconds: 300
-fixture:
-  - bare_remote: origin
-checks:
-  - no_file_changes
-  - no_new_commits

package/templates/base/.agents/skill-evals/sdlc/request-that-fits-no-row/fixture.md

@@ -1 +0,0 @@
-A generated project with a local `origin`.

package/templates/base/.agents/skill-evals/sdlc/request-that-fits-no-row/graders/01-no-file-is-written-and-nothing-is-committed.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No file is written and nothing is committed

package/templates/base/.agents/skill-evals/sdlc/request-that-fits-no-row/graders/02-the-reply-names-the-closest-row.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply says which row of the `/sdlc` table is closest and why it does not fit

package/templates/base/.agents/skill-evals/sdlc/request-that-fits-no-row/graders/03-the-reply-is-short.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply is a few lines long

package/templates/base/.agents/skill-evals/sdlc/request-that-fits-no-row/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 10
-timeout_seconds: 300
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc write a birthday poem for my manager

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/case.yaml

@@ -1,79 +0,0 @@
-schema_version: "1.1"
-name: review-reports-findings-and-never-merges
-expected_outcome: |
-  The branch feat/app-version holds a change with one bug. The agent follows the review playbook: it reviews the diff with a fresh context, reports the finding with its pass, severity, and path:line, and stops at the gate for the code owner. It neither approves nor merges.
-timeout_seconds: 1200
-fixture:
-  - write:
-      path: intent/app-version/intent.md
-      content: |
-        # Intent: app version endpoint
-
-        Author: Joe (Developer Productivity). Status: accepted.
-
-        ## Problem
-
-        Support cannot tell which version of the service a customer reaches. Each ticket about a regression starts with a question about the version, about ten times a week (count from the support queue).
-
-        ## Proposed outcome
-
-        A caller sees the running version at `GET /version`. Out of scope: build metadata.
-
-        ## Affected users and systems
-
-        Support engineers and the service's HTTP API.
-
-        ## Constraints
-
-        No authentication change.
-
-        ## Open questions
-
-        None
-  - write:
-      path: intent/app-version/spec.md
-      content: |
-        # Spec: app version endpoint
-
-        From `intent/app-version/intent.md`. Status: accepted.
-
-        Request: "add a version endpoint". Skills at `0000000`.
-
-        ## Requirements
-
-        1. `GET /version` answers 200 with the JSON body `{"version": "<version in package.json>"}` and no other field.
-
-        ## Design
-
-        `src/version/version.controller.ts` (new) reads the version from `package.json` and returns it.
-
-        ## Policies
-
-        - Brand: Does not apply, no user-visible text.
-        - Security: Does not apply, the endpoint exposes no secret.
-        - Compliance: Does not apply, no personal data.
-        - UX: Does not apply, no interface.
-
-        ## Areas of concern
-
-        None
-
-        ## Open questions
-
-        None
-  - commit: "docs: add app-version intent and spec"
-  - branch: feat/app-version
-  - write:
-      path: src/version/version.controller.ts
-      content: |
-        export function getVersion(): {build: string; version: string} {
-          return {build: "local", version: "1.0.0"};
-        }
-  - commit: "feat: add version controller"
-  - checkout: main
-  - bare_remote: origin
-checks:
-  - reply_matches: '\b(?:Important|Nit)\b'
-  - reply_matches: 'src/version/version\.controller\.ts:\d+'
-  - no_file_changes
-  - no_new_commits

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/fixture.md

@@ -1 +0,0 @@
-The branch `feat/app-version`, which the session did not create, is pushed. Its diff against the default branch contains one bug: `src/version/version.controller.ts` returns a hardcoded version and an extra field, against `intent/app-version/spec.md`. When the session has a pull request host, an open pull request holds that diff.

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/graders/01-findings-have-pass-severity-and-location.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The finding is reported with its pass, its severity, and `path:line`

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/graders/02-no-approval-and-no-merge.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The session neither approves nor merges a pull request

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/graders/03-the-reply-names-the-code-owner-gate.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply names the gate that waits on the code owner

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/graders/04-nothing-is-pushed-or-changed.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The session pushes nothing to the branch `feat/app-version` and changes no file

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/graders/05-the-review-runs-in-a-fresh-context.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The trace shows the review ran in a fresh context: a subagent that was given only the diff, `REVIEW.md`, and the change's spec and plan, or, when no subagent could start, the diff read again from the first line

package/templates/base/.agents/skill-evals/sdlc/review-reports-findings-and-never-merges/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 40
-timeout_seconds: 1200
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc review branch feat/app-version

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/case.yaml

@@ -1,41 +0,0 @@
-schema_version: "1.1"
-name: spec-for-unmerged-intent-stops
-expected_outcome: |
-  The intent intent/dark-mode/intent.md exists only on the branch docs/dark-mode-intent. The agent says in a few lines that the intent is not merged yet, names that branch, and stops. It writes no spec and makes no commit.
-timeout_seconds: 600
-fixture:
-  - branch: docs/dark-mode-intent
-  - write:
-      path: intent/dark-mode/intent.md
-      content: |
-        # Intent: dark mode
-
-        Author: Joe (Developer Productivity). Status: draft.
-
-        ## Problem
-
-        Users who work at night find the white interface tiring.
-
-        ## Proposed outcome
-
-        A dark theme that follows the system setting.
-
-        ## Affected users and systems
-
-        All users of the interface.
-
-        ## Constraints
-
-        None
-
-        ## Open questions
-
-        None
-  - commit: "docs: add dark-mode intent"
-  - checkout: main
-  - bare_remote: origin
-checks:
-  - reply_matches: "docs/dark-mode-intent"
-  - no_file_changes
-  - no_new_commits
-  - no_new_branch

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/fixture.md

@@ -1 +0,0 @@
-`intent/dark-mode/intent.md` exists only on the branch `docs/dark-mode-intent`; the default branch has no `intent/dark-mode/` folder.

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/graders/01-no-spec-is-written.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No `spec.md` is written

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/graders/02-no-commit-is-made.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No new commit is made, nothing is pushed, and no pull request is opened

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/graders/03-the-reply-says-not-merged.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply says that the intent is not merged yet and names the branch `docs/dark-mode-intent`

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/graders/05-the-reply-is-short.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply is at most a few lines and contains no review of the intent's content

package/templates/base/.agents/skill-evals/sdlc/spec-for-unmerged-intent-stops/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 20
-timeout_seconds: 600
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-/sdlc write the spec for intent/dark-mode/intent.md

package/templates/base/.agents/skill-evals/sdlc/statement-without-command-starts-no-work/case.yaml

@@ -1,11 +0,0 @@
-schema_version: "1.1"
-name: statement-without-command-starts-no-work
-expected_outcome: |
-  The message is a statement and no command was typed. The agent replies without creating an intent, a file, or a commit.
-timeout_seconds: 600
-checks:
-  - reply_matches: "/sdlc"
-  - no_file_changes
-  - no_new_commits
-  - no_new_branch
-  - reply_lacks: '(?:^|[\s`("''])/(?:home|tmp|root|Users|var|private)/'

package/templates/base/.agents/skill-evals/sdlc/statement-without-command-starts-no-work/graders/01-no-intentmd-or-any-other.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No `intent.md` or any other file is created or changed

package/templates/base/.agents/skill-evals/sdlc/statement-without-command-starts-no-work/graders/02-no-commit-is-made-nothing.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No commit is made, nothing is pushed, and no pull request is opened

package/templates/base/.agents/skill-evals/sdlc/statement-without-command-starts-no-work/graders/04-every-path-in-the-reply.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-Every path in the reply starts from the project root; none starts from the machine's root

package/templates/base/.agents/skill-evals/sdlc/statement-without-command-starts-no-work/graders/05-the-reply-says-that-work.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The reply says that work starts by sending the request with `/sdlc`

package/templates/base/.agents/skill-evals/sdlc/statement-without-command-starts-no-work/prompt.md

@@ -1,8 +0,0 @@
----
-tags: [skill-creator]
-max_turns: 20
-timeout_seconds: 600
-allowed_tools: [Read, Glob, Grep, Skill, Write, Edit, Bash]
----
-
-The login page is very slow and customers complain

package/templates/base/.agents/skill-evals/sdlc/writes-spec-from-merged-intent/case.yaml

@@ -1,41 +0,0 @@
-schema_version: "1.1"
-name: writes-spec-from-merged-intent
-expected_outcome: |
-  The intent intent/csv-export/intent.md is merged and has no spec. The agent follows the design playbook: it reads the four policies, writes intent/csv-export/spec.md from the template, pushes a branch, and waits at the gate for the product owner. It does not merge.
-timeout_seconds: 900
-fixture:
-  - write:
-      path: intent/csv-export/intent.md
-      content: |
-        # Intent: csv export
-
-        Author: Joe (Developer Productivity). Status: accepted.
-
-        ## Problem
-
-        The accounting team of 12 copies the report table into Excel by hand at every month end, about 20 minutes each (figure from the accounting lead).
-
-        ## Proposed outcome
-
-        One button downloads the open report as a CSV file. Out of scope: PDF and Excel export.
-
-        ## Affected users and systems
-
-        The accounting team and the report page.
-
-        ## Constraints
-
-        Report access rights do not change.
-
-        ## Open questions
-
-        None
-  - commit: "docs: add csv-export intent"
-checks:
-  - file_exists: intent/csv-export/spec.md
-  - file_matches:
-      path: intent/csv-export/spec.md
-      regex: '(?=[\s\S]*^## Requirements$)(?=[\s\S]*^## Design$)(?=[\s\S]*^## Policies$)(?=[\s\S]*^## Areas of concern$)(?=[\s\S]*^## Open questions$)'
-  - commit_touches_only: intent/csv-export/spec.md
-  - branch_matches: "^(?!main$).+"
-  - reply_lacks: '(?:^|[\s`("''])/(?:home|tmp|root|Users|var|private)/'

package/templates/base/.agents/skill-evals/sdlc/writes-spec-from-merged-intent/fixture.md

@@ -1 +0,0 @@
-`intent/csv-export/intent.md` is committed on the default branch, and no `intent/csv-export/spec.md` exists on any branch.

package/templates/base/.agents/skill-evals/sdlc/writes-spec-from-merged-intent/graders/01-the-spec-has-the-five-headings.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-`intent/csv-export/spec.md` exists with exactly the template's five headings: Requirements, Design, Policies, Areas of concern, and Open questions

package/templates/base/.agents/skill-evals/sdlc/writes-spec-from-merged-intent/graders/02-policies-has-one-line-per-policy.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-The spec's Policies section has one line for each of the four policies: brand, security, compliance, and UX

package/templates/base/.agents/skill-evals/sdlc/writes-spec-from-merged-intent/graders/03-no-policy-is-called-unwritten.md

@@ -1,7 +0,0 @@
----
-type: llm
-focus: trace
-weight: 1
----
-
-No line in the spec says that a policy is not written