testeiya 0.3.9 → 0.3.10

package/README.md

@@ -1,73 +1,358 @@
+<p align="center">
+  <img src="https://raw.githubusercontent.com/testomatio/testeiya-app/main/assets/testeiya-cli-logo.png" alt="Testeiya" width="220">
+</p>
+
 # Testeiya
 
+[![npm](https://img.shields.io/npm/v/testeiya)](https://www.npmjs.com/package/testeiya)
+[![licence](https://img.shields.io/npm/l/testeiya)](LICENSE)
+
 An agent, app, and the goddess of testing.
 
-Testeiya is a QA-focused AI agent. It reads and writes manual test cases as markdown, works with automated test code, and talks to [Testomat.io](https://testomat.io) as a test management system. It ships as a desktop app, a web app, and the command-line agent in this repository.
+Testeiya is an autonomous QA agent. It reacts to triggers — a pull request, a new
+issue, a failed test run, a deploy — runs exactly one analysis, and delivers its
+verdict where your team works: a PR comment, a markdown report, or a Testomat.io
+project. There is no interactive mode and no human in the loop. That is the point:
+you wire it into CI once, and every event that should get QA thinking gets it.
 
-This repository holds the parts of Testeiya that shape how the agent thinks:
+It ships as a desktop app, a web app, and the command-line agent in this repository.
 
 | Folder | What it is |
 |---|---|
-| `prompt/` | The system-prompt fragments — the agent's role, rules, tool guidance, Testomat.io operating rules, and the report contract |
-| `skills/` | The manifest of skills the agent can invoke — every folder is fetched from its own upstream repository |
+| `prompt/` | System-prompt fragments: the agent's role, rules, tool guidance, Testomat.io operating rules, and the report contract |
+| `skills/` | The manifest of skills the agent can invoke. Every folder is fetched from its own upstream repository |
 | `src/` | The `testeiya` command-line agent (Node) |
 
-The desktop and web harness — servers, session management, sync, UI — is not open source.
+The desktop and web harness is not open source. That covers the servers, session management, sync, and UI.
+
+## Install
 
-## Install and use
+Requires Node 22.19 or newer, an LLM provider key, and a model.
 
 ```bash
-npx testeiya "Review the manual tests in this folder and list the gaps" \
-  --model openrouter/anthropic/claude-sonnet-5 --output report.md
+npx testeiya doctor
 ```
 
-The agent runs one task and exits — there is no interactive mode. Progress goes to stderr, the report to `--output` (or stdout when you omit it). Exit codes: `0` pass, `1` failed run or negative verdict, `2` bad usage, `130` interrupted — so it drops into CI as-is.
+`doctor` reports which key won, which skills loaded, and whether Testomat.io is reachable — without spending a token.
 
-Requires Node 22.19 or newer, a model, and an LLM provider key.
+The key comes from the environment (`OPENROUTER_API_KEY`, `ANTHROPIC_API_KEY`,
+`OPENAI_API_KEY`, `GEMINI_API_KEY`), from `~/.testeiya/.env`, or from
+`~/.testeiya/auth.json`. That last file is the one the desktop app's Settings
+dialog writes, so configuring it once covers both.
 
-There is no default model. Name one with `--model <provider>/<id>` or `TESTEIYA_MODEL`; a run that names none exits `2` rather than spending your money on a model nobody chose.
-
-The key comes from the environment (`OPENROUTER_API_KEY` and friends), `~/.testeiya/.env`, or `~/.testeiya/auth.json` — the same file the desktop app's Settings dialog writes, so configuring it once covers both.
+There is no default model. Name one with `--model <provider>/<id>` or
+`TESTEIYA_MODEL`. CI usually has neither set, so a run that resolves no model
+exits `2` rather than picking one for you.
 
 ```bash
-testeiya --help                      # every flag
 export TESTEIYA_MODEL=openrouter/anthropic/claude-sonnet-5
-testeiya "<task>"
-cat task.md | testeiya --output report.md
+testeiya models anthropic   # list what your key can reach
+```
+
+## How it runs
+
+```bash
+testeiya task "<task>"     # run one task, deliver a report, exit
+testeiya ask "<question>"  # answer a question, no report
+testeiya sessions          # list saved sessions for this folder
+```
+
+The agent runs one task and exits. Progress goes to stderr, so a run drops into CI as-is:
+
+| Exit code | Meaning |
+|---|---|
+| `0` | pass, or a positive verdict from the agent |
+| `1` | failed run, or a negative verdict — findings that need a human |
+| `2` | bad usage |
+| `130` | interrupted |
+
+A task can come from stdin too:
+
+```bash
+cat issue-42.md | testeiya task --output report.md
+```
+
+Every command takes `--json` for machine-readable output of the run envelope
+(verdict, reason, tokens, session id). `testeiya --help` lists all options;
+`testeiya help` is the full guide.
+
+## Scenarios
+
+Each scenario below is a single unattended run. Nobody answers questions; the
+agent reads what the trigger gives it and acts on its own judgement. If its
+verdict is negative, the job fails — that failure *is* the signal.
+
+### Grill a pull request
+
+Every PR gets a QA review before merge. The agent loads the branch's diff,
+analyzes it through the `qa-thinking` skill — edge cases, negative flows,
+abuses, data-consistency risks — and posts the findings as a PR comment:
+
+```bash
+git fetch origin "$PR_BRANCH"
+git checkout "$PR_BRANCH"
+testeiya task "Review this pull request as a QA engineer. What could go wrong?" \
+  --output gh:pr-comment --output review.json
+```
+
+Exit code `1` means the agent found real risk, so you can gate the merge on it.
+
+### Write test cases from a pull request
+
+Same trigger, different deliverable: the `qa-write-test-cases` skill turns the
+change into test cases in Testomat.io markdown format, written as a build
+artifact ready to commit or import:
+
+```bash
+testeiya task "Create test cases covering the changes in this pull request" \
+  --output testcases/
+```
+
+With `TESTOMATIO` and the project id set, add
+`sync them to Testomat.io` to the task and they land straight in the TMS via
+`sync-test-cases-with-tms`.
+
+### Create test cases from a new issue
+
+A requirements text arrives from the issue tracker — piped in, no human
+summarizing it first. The agent writes a checklist and test cases from it:
+
+```bash
+gh issue view 57 --json title,body -q '.title + "\n\n" + .body' \
+| testeiya task --output testcases/issue-57.md
 ```
 
-Set `TESTOMATIO` to a project API key and the agent can read and write that project's tests, suites, runs and plans through `check-tests` and the REST API. Add the project id as well and it also gets the Testomat.io tools:
+### Explore a deployed app with Explorbot
+
+After a deploy, point [Explorbot](https://github.com/testomatio/explorbot) — the
+autonomous browser-testing CLI the agent drives through the `explorbot-*`
+skills — at the staging URL. It researches, plans, and tests the live app in
+its own browser, then reports what broke:
 
 ```bash
-TESTOMATIO=tstmt_xxx testeiya --project my-project "Which suites have no tests?"
+testeiya task "Run explorbot against https://staging.example.com, max 10 tests, report failures" \
+  --output explorbot-report.md
 ```
 
-The id can come from `--project` or `TESTOMATIO_PROJECT_ID`. The MCP server needs it: a token alone does not tell it which project to talk to.
+### Repair failing tests after a red build
+
+On a failed CodeceptJS run, the `ci-fix-tests` skill attempts safe fixes only —
+locator drift, missing waits — reruns just the failing scenarios, rolls back any
+edit that did not help, and always writes `output/ci-fix.md` for the next job:
+
+```bash
+testeiya task "Fix the failed CodeceptJS tests using ci-fix-tests. No refactors." \
+  --no-session
+```
+
+### Audit the test suite on a schedule
+
+Nightly, the agent walks the Testomat.io project and reports gaps — suites with
+no tests, cases gone stale against the current code, coverage holes:
+
+```bash
+TESTOMATIO=tstmt_xxx testeiya task \
+  "Audit this project: which suites have no automated tests, and which manual cases look automatable?" \
+  --output audits/$(date +%F).md
+```
+
+## CI setup
+
+Testeiya needs three things in any CI system: a provider key, a model name, and
+the checkout of whatever the task reads. Everything else is standard.
+
+### GitHub Actions
+
+GitHub runners ship the [GitHub CLI](https://cli.github.com), so posting back to
+the PR is one flag. `GITHUB_TOKEN` authenticates it; keep the LLM key in
+repository secrets.
+
+```yaml
+name: qa-review
+on:
+  pull_request:
+    types: [opened, synchronize]
+
+permissions:
+  pull-requests: write
+  contents: read
+
+jobs:
+  grill:
+    runs-on: ubuntu-latest
+    env:
+      GH_TOKEN: ${{ github.token }}
+      OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }}
+      TESTEIYA_MODEL: openrouter/anthropic/claude-sonnet-5
+      # optional, for scenarios that touch Testomat.io
+      TESTOMATIO: ${{ secrets.TESTOMATIO }}
+      TESTOMATIO_PROJECT_ID: my-project
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - uses: actions/setup-node@v4
+        with:
+          node-version: 22
+      - run: npx testeiya@latest doctor
+      - run: |
+          npx testeiya@latest task \
+            "Review this pull request as a QA engineer. What could go wrong?" \
+            --output gh:pr-comment --output review.json
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: review
+          path: review.json
+```
+
+The same shape works for the other scenarios: trigger on `issues` and pipe the
+body in for test-case generation, or run on a `schedule` for the nightly audit.
+
+### GitLab CI
+
+GitLab runners do not ship `gh`, so the report goes to job artifacts — visible
+in the merge request pipeline page. Exit code `1` fails the job and blocks the
+merge when you mark it required.
+
+```yaml
+qa-review:
+  image: node:22
+  rules:
+    - if: $CI_PIPELINE_SOURCE == "merge_request_event"
+  variables:
+    TESTEIYA_MODEL: openrouter/anthropic/claude-sonnet-5
+    TESTOMATIO_PROJECT_ID: my-project
+  script:
+    - git fetch origin $CI_MERGE_REQUEST_TARGET_BRANCH_NAME
+    - npx testeiya@latest doctor
+    - npx testeiya@latest task
+        "Review this merge request as a QA engineer. What could go wrong?"
+        --output review/report.md --output review/envelope.json
+  artifacts:
+    when: always
+    paths:
+      - review/
+  parallel:
+    # secrets go to Settings > CI/CD > Variables:
+    #   OPENROUTER_API_KEY (masked), TESTOMATIO (masked, protected)
+```
+
+For posting back to the merge request, hand the envelope's report to GitLab's
+API in a follow-up step, or install `gh` plus a token mirror if the project is
+also on GitHub.
+
+### Anywhere else
+
+Any scheduler that can run a container works — the contract is just stdin,
+stdout, and an exit code:
+
+```bash
+echo "Audit the checkout suite for gaps" | testeiya task --output audit.md
+case $? in
+  0) echo "clean" ;;
+  1) echo "findings need review" ;;
+esac
+```
+
+## Where the report goes
+
+`--output` names a destination and can repeat. Without one the report goes to stdout.
+
+```bash
+testeiya task "<task>" --output report.md --output run.json --output gh:pr-comment
+```
+
+| Destination | What happens |
+|---|---|
+| `report.md` | the agent writes the report there |
+| `run.json` | the run envelope: verdict, reason, report, tokens, session id |
+| `gh:pr-comment` | posted on this branch's pull request |
+| `gh:pr#123` | posted on that pull request |
+
+Posting uses the GitHub CLI and resolves the PR number from `GITHUB_EVENT_PATH`,
+`GITHUB_REF`, or `gh pr view` — so inside Actions it just works. Every
+destination is checked before the run starts, so a missing `gh` costs no tokens.
+
+## Sessions
+
+Runs are saved under `~/.testeiya`, so a follow-up picks up where the last one
+stopped. A resumed run reuses its session's model.
+
+```bash
+testeiya task "Review the checkout suite" --name checkout-review --output report.md
+testeiya task "Now write the missing cases" -c
+testeiya sessions
+testeiya task "<task>" --resume <id>
+```
+
+Pass `--name` to label a session and `--no-session` to save nothing. In CI,
+`--no-session` keeps runners stateless unless you deliberately chain `-c` steps.
+
+## Testomat.io
+
+Set `TESTOMATIO` to a project API key. The agent can then read and write that
+project's tests, suites, runs and plans through `check-tests` and the REST API.
+
+```bash
+TESTOMATIO=tstmt_xxx testeiya task --project my-project "Which suites have no tests?"
+```
+
+Add the project id and the agent also gets the Testomat.io MCP tools. The id
+comes from `--project` or from `TESTOMATIO_PROJECT_ID`; the MCP server needs it,
+because a token alone does not say which project to talk to. `TESTOMATIO_URL`
+points at a self-hosted instance.
 
 ## Skills
 
-A skill is a folder with a `SKILL.md`. The agent loads them all and invokes the ones a task calls for.
+A skill is a folder with a `SKILL.md`. The agent loads them all and invokes the
+ones a task calls for — you never pick a skill explicitly, the task phrasing
+does.
+
+What ships today, vendored from upstream repositories:
+
+| Category | Skills |
+|---|---|
+| QA process | `testing-workflow`, `qa-explain-behavior`, `qa-lead-strategy-advisor` |
+| Test management | `qa-write-test-cases`, `sync-test-cases-with-tms`, `qa-sprint-report-by-testomatio`, `qa-split-testing-levels-pyramid`, `qa-thinking` |
+| Test automation | `automate-manual-test-cases`, `setup-change-aware-pr-testing`, `run-tests-with-testomatio-reporter`, `setup-ci-automation`, `debug-fix-failed-flaky-autotests`, `qa-data-seeder`, `qa-automation-test-consolidation`, `testomat-allure-adapter` |
+| Explorbot | `explorbot-fundamentals`, `explorbot-setup`, `explorbot-plan` |
+| Playwright | `playwright-best-practices`, `playwright-cli` |
+| CodeceptJS | writing, refactoring, debugging, run analysis, migrations from Cypress/TestCafe/Selenium/Protractor, `ci-fix-tests` |
 
-Every skill here is vendored from its own upstream repository, declared in `skills/skills.yaml` and pinned to a commit in `skills/skills.lock.json`. The vendored folders are deliberately **not** committed — they belong to their authors, under their own licences. A clone of this repository has the manifest and nothing else; the tree is fetched by the upstream release tooling, which is where the desktop app gets its full set.
+Sources are declared in `skills/skills.yaml` and pinned in
+`skills/skills.lock.json`. The vendored folders are deliberately not committed —
+they belong to their authors, under their own licences. A clone has the manifest
+and nothing else; `node scripts/vendor-skills.js` fills the tree. Every release
+runs it, so the published `testeiya` package ships each skill as current on
+release day.
 
-The published `testeiya` package therefore ships no skills of its own: `skillsOverride` in `src/session.ts` keeps only what is found under the bundled tree, so a fresh `npx testeiya` run has none until that tree is filled. To add your own, point `additionalSkillPaths` at your folder — [EXTENDING.md](EXTENDING.md) covers both hooks. The first-party skills the desktop app bundles are written against tools only that harness has, so they live with it, in the private repository.
+`skillsOverride` in `src/session.ts` keeps only what is found under that tree,
+so an arbitrary clone cannot hand the model its own skills. To add yours, point
+`additionalSkillPaths` at your folder. [EXTENDING.md](EXTENDING.md) covers both hooks.
 
-To propose a new source, add its line to `skills/skills.yaml` — see [CONTRIBUTING.md](CONTRIBUTING.md).
+To propose a new source, add its line to `skills/skills.yaml`. See [CONTRIBUTING.md](CONTRIBUTING.md).
 
 ## Building your own agent
 
-The CLI is a thin composition over [pi](https://pi.dev): under 850 lines wiring the SDK to the prompt and skills here. A fork can add pi extensions and custom tools, or swap the one-shot run loop for pi's full interactive TUI — [EXTENDING.md](EXTENDING.md) walks through both.
+The CLI is a thin composition over [pi](https://pi.dev): about 1,500 lines wiring
+the SDK to the prompt and skills here. A fork can add pi extensions and custom
+tools, or swap the one-shot run loop for pi's full interactive TUI.
+[EXTENDING.md](EXTENDING.md) walks through both.
 
 ## Contributing
 
-Prompt wording is exactly what an outside contributor can improve, and a change to it changes how the agent behaves for everyone. Read [CONTRIBUTING.md](CONTRIBUTING.md) first — it covers what belongs here and what belongs upstream, in the repository that owns a given skill.
+Prompt wording is exactly what an outside contributor can improve, and a change
+to it changes how the agent behaves for everyone. Read
+[CONTRIBUTING.md](CONTRIBUTING.md) first. It covers what belongs here and what
+belongs upstream, in the repository that owns a given skill.
 
 ## Issues
 
 This repository is also the public issue tracker for both surfaces:
 
-- **Testeiya Desktop app** — the packaged desktop application
-- **Testeiya CLI** — the command-line agent in `src/`
+- Testeiya Desktop app, the packaged desktop application
+- Testeiya CLI, the command-line agent in `src/`
 
 ## Licence
 

package/dist/prompt/index.js

@@ -1,6 +1,6 @@
 import { getSystemPrompt } from "./system-prompt.js";
 import { testomatioTms, testomatioConnection, testomatioNotConnected, projectSettings, } from "./testomatio.js";
-import { nonInteractive, reportOutput } from "./print.js";
+import { briefAnswer, nonInteractive, reportOutput } from "./print.js";
 import { contextPromptSection, } from "./context.js";
 export function buildSystemPrompt(options) {
     const mode = options?.mode ?? "tui";
@@ -15,6 +15,8 @@ export function buildSystemPrompt(options) {
             interactive,
             toolBullets: options?.toolBullets,
             rules: options?.rules,
+            connectedClis: options?.connectedClis,
+            supportedMcps: options?.supportedMcps,
         }),
     ];
     const tokenSlugs = options?.tokens ? Object.keys(options.tokens) : [];
@@ -48,7 +50,9 @@ export function buildSystemPrompt(options) {
     else if (mode === "web" || mode === "print") {
         parts.push(testomatioNotConnected());
     }
-    // The report contract goes last so it is the final instruction the model reads.
+    // The answer contract goes last so it is the final instruction the model reads.
+    if (options?.brief)
+        parts.push(briefAnswer);
     if (options?.outputFile)
         parts.push(reportOutput(options.outputFile));
     return parts.join("\n\n");

package/dist/prompt/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../../prompt/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,OAAO,EACL,aAAa,EACb,oBAAoB,EACpB,sBAAsB,EACtB,eAAe,GAEhB,MAAM,iBAAiB,CAAC;AACzB,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAC1D,OAAO,EACL,oBAAoB,GAGrB,MAAM,cAAc,CAAC;AAGtB,MAAM,UAAU,iBAAiB,CAAC,OAA6B;IAC7D,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,IAAI,KAAK,CAAC;IACpC,MAAM,WAAW,GAAG,IAAI,KAAK,OAAO,CAAC;IAErC,+EAA+E;IAC/E,gFAAgF;IAChF,6EAA6E;IAC7E,oEAAoE;IACpE,8CAA8C;IAC9C,MAAM,KAAK,GAAa;QACtB,eAAe,CAAC,OAAO,EAAE,GAAG,EAAE;YAC5B,WAAW;YACX,WAAW,EAAE,OAAO,EAAE,WAAW;YACjC,KAAK,EAAE,OAAO,EAAE,KAAK;SACtB,CAAC;KACH,CAAC;IAEF,MAAM,UAAU,GAAG,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACtE,yEAAyE;IACzE,MAAM,cAAc,GAAG,OAAO,EAAE,UAAU,EAAE,cAAc,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC;IACpF,MAAM,GAAG,GAAG,OAAO,EAAE,GAAG,IAAI,YAAY,CAAC;IACzC,6EAA6E;IAC7E,6EAA6E;IAC7E,qEAAqE;IACrE,IAAI,cAAc,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5E,IAAI,IAAI,KAAK,OAAO;QAAE,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;IACjD,8EAA8E;IAC9E,2EAA2E;IAC3E,yEAAyE;IACzE,KAAK,MAAM,OAAO,IAAI,OAAO,EAAE,QAAQ,IAAI,EAAE;QAAE,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAEnE,IAAI,OAAO,EAAE,aAAa,EAAE,CAAC;QAC3B,KAAK,CAAC,IAAI,CAAC,8BAA8B,OAAO,CAAC,aAAa,EAAE,CAAC,CAAC;IACpE,CAAC;IACD,IAAI,OAAO,EAAE,cAAc,EAAE,MAAM,IAAI,OAAO,EAAE,cAAc,EAAE,MAAM,EAAE,CAAC;QACvE,KAAK,CAAC,IAAI,CACR,oBAAoB,CAAC,OAAO,CAAC,cAAc,IAAI,EAAE,EAAE,OAAO,CAAC,cAAc,IAAI,EAAE,CAAC,CACjF,CAAC;IACJ,CAAC;IACD,IAAI,OAAO,EAAE,WAAW,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC;IACnD,CAAC;IACD,IAAI,cAAc,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5C,KAAK,CAAC,IAAI,CAAC,oBAAoB,CAAC,UAAU,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC,CAAC;IAC9F,CAAC;SAAM,IAAI,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,OAAO,EAAE,CAAC;QAC9C,KAAK,CAAC,IAAI,CAAC,sBAAsB,EAAE,CAAC,CAAC;IACvC,CAAC;IAED,gFAAgF;IAChF,IAAI,OAAO,EAAE,UAAU;QAAE,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC;IAEtE,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC5B,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../prompt/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,OAAO,EACL,aAAa,EACb,oBAAoB,EACpB,sBAAsB,EACtB,eAAe,GAEhB,MAAM,iBAAiB,CAAC;AACzB,OAAO,EAAE,WAAW,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AACvE,OAAO,EACL,oBAAoB,GAGrB,MAAM,cAAc,CAAC;AAGtB,MAAM,UAAU,iBAAiB,CAAC,OAA6B;IAC7D,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,IAAI,KAAK,CAAC;IACpC,MAAM,WAAW,GAAG,IAAI,KAAK,OAAO,CAAC;IAErC,+EAA+E;IAC/E,gFAAgF;IAChF,6EAA6E;IAC7E,oEAAoE;IACpE,8CAA8C;IAC9C,MAAM,KAAK,GAAa;QACtB,eAAe,CAAC,OAAO,EAAE,GAAG,EAAE;YAC5B,WAAW;YACX,WAAW,EAAE,OAAO,EAAE,WAAW;YACjC,KAAK,EAAE,OAAO,EAAE,KAAK;YACrB,aAAa,EAAE,OAAO,EAAE,aAAa;YACrC,aAAa,EAAE,OAAO,EAAE,aAAa;SACtC,CAAC;KACH,CAAC;IAEF,MAAM,UAAU,GAAG,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACtE,yEAAyE;IACzE,MAAM,cAAc,GAAG,OAAO,EAAE,UAAU,EAAE,cAAc,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC;IACpF,MAAM,GAAG,GAAG,OAAO,EAAE,GAAG,IAAI,YAAY,CAAC;IACzC,6EAA6E;IAC7E,6EAA6E;IAC7E,qEAAqE;IACrE,IAAI,cAAc,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5E,IAAI,IAAI,KAAK,OAAO;QAAE,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;IACjD,8EAA8E;IAC9E,2EAA2E;IAC3E,yEAAyE;IACzE,KAAK,MAAM,OAAO,IAAI,OAAO,EAAE,QAAQ,IAAI,EAAE;QAAE,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAEnE,IAAI,OAAO,EAAE,aAAa,EAAE,CAAC;QAC3B,KAAK,CAAC,IAAI,CAAC,8BAA8B,OAAO,CAAC,aAAa,EAAE,CAAC,CAAC;IACpE,CAAC;IACD,IAAI,OAAO,EAAE,cAAc,EAAE,MAAM,IAAI,OAAO,EAAE,cAAc,EAAE,MAAM,EAAE,CAAC;QACvE,KAAK,CAAC,IAAI,CACR,oBAAoB,CAAC,OAAO,CAAC,cAAc,IAAI,EAAE,EAAE,OAAO,CAAC,cAAc,IAAI,EAAE,CAAC,CACjF,CAAC;IACJ,CAAC;IACD,IAAI,OAAO,EAAE,WAAW,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC;IACnD,CAAC;IACD,IAAI,cAAc,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5C,KAAK,CAAC,IAAI,CAAC,oBAAoB,CAAC,UAAU,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC,CAAC;IAC9F,CAAC;SAAM,IAAI,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,OAAO,EAAE,CAAC;QAC9C,KAAK,CAAC,IAAI,CAAC,sBAAsB,EAAE,CAAC,CAAC;IACvC,CAAC;IAED,gFAAgF;IAChF,IAAI,OAAO,EAAE,KAAK;QAAE,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;IAC5C,IAAI,OAAO,EAAE,UAAU;QAAE,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC;IAEtE,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC5B,CAAC"}

package/dist/prompt/print.js

@@ -19,4 +19,12 @@ export function reportOutput(path) {
 </final-report>
 `;
 }
+export const briefAnswer = dedent `
+<answer>
+  * You were asked a question, not given a task. Answer it.
+  * Lead with the answer in one line, then the evidence you checked.
+  * A few sentences. No report file, no headings, no plan.
+  * Say plainly when what you found does not settle the question.
+</answer>
+`;
 //# sourceMappingURL=print.js.map

package/dist/prompt/print.js.map

@@ -1 +1 @@
-{"version":3,"file":"print.js","sourceRoot":"","sources":["../../prompt/print.ts"],"names":[],"mappings":"AAAA,OAAO,MAAM,MAAM,QAAQ,CAAC;AAE5B,MAAM,CAAC,MAAM,cAAc,GAAG,MAAM,CAAA;;;;;;;;;CASnC,CAAC;AAEF,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,OAAO,MAAM,CAAA;;4CAE6B,IAAI;;;;;CAK/C,CAAC;AACF,CAAC"}
+{"version":3,"file":"print.js","sourceRoot":"","sources":["../../prompt/print.ts"],"names":[],"mappings":"AAAA,OAAO,MAAM,MAAM,QAAQ,CAAC;AAE5B,MAAM,CAAC,MAAM,cAAc,GAAG,MAAM,CAAA;;;;;;;;;CASnC,CAAC;AAEF,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,OAAO,MAAM,CAAA;;4CAE6B,IAAI;;;;;CAK/C,CAAC;AACF,CAAC;AAED,MAAM,CAAC,MAAM,WAAW,GAAG,MAAM,CAAA;;;;;;;CAOhC,CAAC"}

package/dist/prompt/system-prompt.js

@@ -9,6 +9,8 @@ export function getSystemPrompt(cwd, options) {
     let extraRules = "";
     for (const rule of options?.rules ?? [])
         extraRules += `    * ${rule}\n`;
+    let cliList = (options?.connectedClis ?? []).join(", ") || "none connected yet";
+    let mcpList = (options?.supportedMcps ?? []).join(", ") || "none connected yet";
     return dedent `
   <role>
     You are Testeiya, an AI agent that helps with QA tasks.
@@ -87,6 +89,16 @@ export function getSystemPrompt(cwd, options) {
 
   ${tools({ extra: options?.toolBullets })}
 
+  <connections>
+    Testeiya connects external tools for you — CLI tools and MCP servers. This session has:
+
+    * **Connected CLIs:** ${cliList}
+    * **Supported MCP servers:** ${mcpList}
+
+    * **Missing Connection — Ask, Never Improvise:** When a task needs a tool that is not in the lists above, STOP and ask the user to connect it in Testeiya (Settings → Connections). Never reach the service sideways: no raw REST/GraphQL calls against its API, no scraping credentials from dotfiles or env dumps, no installing binaries on your own.
+    * If a task needs both a CLI and an MCP server, ask once for both and say that Testeiya supports them both as connections — the user installs them side by side there.
+  </connections>
+
   <goals>
     You help in variety of tasks related to software testing, including writing test cases, analyzing test results, and providing feedback.
     You have skills to perform QA tasks on user demand.
@@ -98,10 +110,8 @@ export function getSystemPrompt(cwd, options) {
     * Assisting with manual test execution
     * Manage tests with TMS Testomat.io
     * Setting up CI pipelines for continuous testing
-    * Analyzing requirements
-    * Automating manual test cases
-    * Launching automated exploratory tests via explorbot
-    * Analyzing test results and providing feedback
+    * Analyzing requirements and issues
+    * Analyzing test results
 
     This is what you can do.
   </goals>

package/dist/prompt/system-prompt.js.map

@@ -1 +1 @@
-{"version":3,"file":"system-prompt.js","sourceRoot":"","sources":["../../prompt/system-prompt.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,YAAY,CAAC;AACnC,OAAO,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAC;AAE/C,OAAO,MAAM,MAAM,QAAQ,CAAC;AAE5B,MAAM,UAAU,eAAe,CAAC,GAAY,EAAE,OAAuB;IACnE,MAAM,WAAW,GAAG,OAAO,EAAE,WAAW,IAAI,IAAI,CAAC;IACjD,IAAI,mBAAmB,GAAG,qCAAqC,CAAC;IAChE,IAAI,CAAC,WAAW;QAAE,mBAAmB,GAAG,uCAAuC,CAAC;IAChF,IAAI,UAAU,GAAG,EAAE,CAAC;IACpB,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,KAAK,IAAI,EAAE;QAAE,UAAU,IAAI,SAAS,IAAI,IAAI,CAAC;IACzE,OAAO,MAAM,CAAA;;;;;;;;;;;qBAWM,GAAG,IAAI,OAAO,CAAC,GAAG,EAAE;oEAC2B,iBAAiB;yEACZ,iBAAiB;uCACnD,iBAAiB;wEACgB,iBAAiB;0BAC/D,iBAAiB,0GAA0G,iBAAiB,0DAA0D,iBAAiB;;yHAExH,iBAAiB;;;;2CAI/F,iBAAiB;UAClD,iBAAiB;;;;;;;gWAOqU,iBAAiB;2FACtR,iBAAiB,sJAAsJ,iBAAiB;;6EAEtM,iBAAiB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;MAyCxF,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;;IAExC,KAAK,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,CAAC;;;;;;;;;;;;;;;;;;;;;;;;;gCAyBV,WAAW,CAAC,CAAC,CAAC,sEAAsE,CAAC,CAAC,CAAC,6EAA6E;;;;;;;;;;;;;;;;;;;EAmBlM,UAAU,6HAA6H,mBAAmB;;;;;gFAK5E,iBAAiB;;;;GAI9F,CAAC;AACJ,CAAC;AAWD;;;;;;;;;;;;;;;;;;;;EAoBE"}
+{"version":3,"file":"system-prompt.js","sourceRoot":"","sources":["../../prompt/system-prompt.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,YAAY,CAAC;AACnC,OAAO,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAC;AAE/C,OAAO,MAAM,MAAM,QAAQ,CAAC;AAE5B,MAAM,UAAU,eAAe,CAAC,GAAY,EAAE,OAAuB;IACnE,MAAM,WAAW,GAAG,OAAO,EAAE,WAAW,IAAI,IAAI,CAAC;IACjD,IAAI,mBAAmB,GAAG,qCAAqC,CAAC;IAChE,IAAI,CAAC,WAAW;QAAE,mBAAmB,GAAG,uCAAuC,CAAC;IAChF,IAAI,UAAU,GAAG,EAAE,CAAC;IACpB,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,KAAK,IAAI,EAAE;QAAE,UAAU,IAAI,SAAS,IAAI,IAAI,CAAC;IACzE,IAAI,OAAO,GAAG,CAAC,OAAO,EAAE,aAAa,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,oBAAoB,CAAC;IAChF,IAAI,OAAO,GAAG,CAAC,OAAO,EAAE,aAAa,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,oBAAoB,CAAC;IAChF,OAAO,MAAM,CAAA;;;;;;;;;;;qBAWM,GAAG,IAAI,OAAO,CAAC,GAAG,EAAE;oEAC2B,iBAAiB;yEACZ,iBAAiB;uCACnD,iBAAiB;wEACgB,iBAAiB;0BAC/D,iBAAiB,0GAA0G,iBAAiB,0DAA0D,iBAAiB;;yHAExH,iBAAiB;;;;2CAI/F,iBAAiB;UAClD,iBAAiB;;;;;;;gWAOqU,iBAAiB;2FACtR,iBAAiB,sJAAsJ,iBAAiB;;6EAEtM,iBAAiB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;MAyCxF,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;;IAExC,KAAK,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,CAAC;;;;;4BAKd,OAAO;mCACA,OAAO;;;;;;;;;;;;;;;;;;;;;;;;;;;gCA2BV,WAAW,CAAC,CAAC,CAAC,sEAAsE,CAAC,CAAC,CAAC,6EAA6E;;;;;;;;;;;;;;;;;;;EAmBlM,UAAU,6HAA6H,mBAAmB;;;;;gFAK5E,iBAAiB;;;;GAI9F,CAAC;AACJ,CAAC;AAeD;;;;;;;;;;;;;;;;;;;;EAoBE"}

package/dist/prompt/testomatio.js

@@ -32,7 +32,7 @@ const tmsViaDirectTools = dedent `
   | Runs, testruns, plans, labels, issues, analytics | MCP \`testomatio-*\` tools |
   | Creating/updating tests or suites | edit the markdown file, then \`npx check-tests push\` |
   | Creating/updating runs, linking issues | MCP create/update tools |
-  | Launching automated or mixed Run remotely on CI | Use npx @testomatio/reporter run --remote <profile>, raise error if no CI profiles configured
+  | Launching automated or mixed Run remotely on CI | Use npx -y @testomatio/reporter@latest run --remote <profile>, raise error if no CI profiles configured
   | Running automated Run | Use local test runner with testomatio reporter attached to report to current project
 
   ---
@@ -69,7 +69,7 @@ const tmsViaProxy = dedent `
   | Runs, testruns, plans, labels, issues, analytics | \`mcp({ search: … })\`, then \`mcp({ tool: … })\` |
   | Creating/updating tests or suites | edit the markdown file, then \`npx check-tests push\` |
   | Creating/updating runs, linking issues | the matching \`mcp\` create/update operation |
-  | Launching automated or mixed Run remotely on CI | Use npx @testomatio/reporter run --remote <profile>, raise error if no CI profiles configured |
+  | Launching automated or mixed Run remotely on CI | Use npx -y @testomatio/reporter@latest run --remote <profile>, raise error if no CI profiles configured |
   | Running automated Run | Use local test runner with testomatio reporter attached to report to current project |
 
   ---
@@ -101,7 +101,7 @@ const tmsViaCli = dedent `
   | Suite hierarchy / structure / which files exist | \`ls\` / \`find\` |
   | Runs, testruns, plans, labels, issues, analytics | \`curl\` against \`/api/v2\` |
   | Creating/updating tests or suites | edit the markdown file, then \`npx check-tests push\` |
-  | Launching automated or mixed Run remotely on CI | Use npx @testomatio/reporter run --remote <profile>, raise error if no CI profiles configured |
+  | Launching automated or mixed Run remotely on CI | Use npx -y @testomatio/reporter@latest run --remote <profile>, raise error if no CI profiles configured |
   | Running automated Run | Use local test runner with testomatio reporter attached to report to current project |
 
   ---

package/dist/prompt/tools.js

@@ -9,11 +9,19 @@ import dedent from 'dedent';
  */
 export function tools(options) {
     const extra = (options?.extra ?? []).join('');
-    let github = '    - GitHub: issues and projects through MCP.';
-    if (hasGh()) {
-        github +=
+    let routing = '    - GitHub: issues and projects through MCP.';
+    if (onPath('gh')) {
+        routing +=
             '\n    - Pull requests and repositories: `gh` in `bash`. When a PR question needs issue detail, query the issue through MCP — never infer it from a PR title.';
     }
+    if (onPath('acli')) {
+        routing +=
+            '\n    - Jira in bulk (many issues at once), boards, sprints and filters: `acli` in `bash`, per the `atlassian` skill. Confluence and attachments have no `acli` command and stay on MCP.';
+    }
+    if (onPath('glab')) {
+        routing +=
+            '\n    - GitLab releases, labels, milestones, snippets, CI artifacts, and any endpoint with no MCP tool: `glab` in `bash`. Issues, merge requests, pipelines and code search stay on MCP, which covers them well.';
+    }
     return dedent `
 <available-tools>
   You have these tools available:
@@ -25,7 +33,7 @@ ${extra}  * **Internal Skills:** For QA-related tasks, check available skills fi
   * **External Integrations:** Use MCP tools when they provide superior data or specialized integrations.
     - Primary MCP (Testomat.io): Use Testomat.io MCP tools to extend functionality.
     - Secondary MCP (Jira, GitHub, etc.): Invoke only when user explicitly asks or when remote context exploration is required.
-${github}
+${routing}
 
   <tool-governance>
     * **Prioritize Specificity:** Use \`read\` instead of \`cat\`, and \`edit\` instead of \`sed\`. Dedicated tools provide better error tracking and safety.
@@ -36,12 +44,12 @@ ${github}
 </available-tools>
 `;
 }
-/** Is the `gh` CLI on PATH? Never tell the agent to use a binary this machine lacks. */
-function hasGh() {
+/** Is this CLI on PATH? Never tell the agent to use a binary this machine lacks. */
+function onPath(bin) {
     for (const dir of (process.env.PATH ?? '').split(delimiter)) {
         if (!dir)
             continue;
-        if (existsSync(join(dir, 'gh')) || existsSync(join(dir, 'gh.exe')))
+        if (existsSync(join(dir, bin)) || existsSync(join(dir, `${bin}.exe`)))
             return true;
     }
     return false;

package/dist/prompt/tools.js.map

@@ -1 +1 @@
-{"version":3,"file":"tools.js","sourceRoot":"","sources":["../../prompt/tools.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACrC,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAE5C,OAAO,MAAM,MAAM,QAAQ,CAAC;AAE5B;;;;;GAKG;AACH,MAAM,UAAU,KAAK,CAAC,OAA8B;IAClD,MAAM,KAAK,GAAG,CAAC,OAAO,EAAE,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC9C,IAAI,MAAM,GAAG,gDAAgD,CAAC;IAC9D,IAAI,KAAK,EAAE,EAAE,CAAC;QACZ,MAAM;YACJ,8JAA8J,CAAC;IACnK,CAAC;IACD,OAAO,MAAM,CAAA;;;;;;;EAOb,KAAK;;;;EAIL,MAAM;;;;;;;;;CASP,CAAC;AACF,CAAC;AAED,wFAAwF;AACxF,SAAS,KAAK;IACZ,KAAK,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;QAC5D,IAAI,CAAC,GAAG;YAAE,SAAS;QACnB,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;YAAE,OAAO,IAAI,CAAC;IAClF,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC"}
+{"version":3,"file":"tools.js","sourceRoot":"","sources":["../../prompt/tools.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACrC,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAE5C,OAAO,MAAM,MAAM,QAAQ,CAAC;AAE5B;;;;;GAKG;AACH,MAAM,UAAU,KAAK,CAAC,OAA8B;IAClD,MAAM,KAAK,GAAG,CAAC,OAAO,EAAE,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC9C,IAAI,OAAO,GAAG,gDAAgD,CAAC;IAC/D,IAAI,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;QACjB,OAAO;YACL,8JAA8J,CAAC;IACnK,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;QACnB,OAAO;YACL,0LAA0L,CAAC;IAC/L,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;QACnB,OAAO;YACL,kNAAkN,CAAC;IACvN,CAAC;IACD,OAAO,MAAM,CAAA;;;;;;;EAOb,KAAK;;;;EAIL,OAAO;;;;;;;;;CASR,CAAC;AACF,CAAC;AAED,oFAAoF;AACpF,SAAS,MAAM,CAAC,GAAW;IACzB,KAAK,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;QAC5D,IAAI,CAAC,GAAG;YAAE,SAAS;QACnB,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,GAAG,MAAM,CAAC,CAAC;YAAE,OAAO,IAAI,CAAC;IACrF,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC"}