Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions .claude-plugin/marketplace.json
Original file line number Diff line number Diff line change
Expand Up @@ -78,6 +78,29 @@
"verified"
]
},
{
"name": "rf-mind",
"source": "./plugins/rf-mind",
"description": "robotframework-mind skill for exact lookups in a Robot Framework test repo: tickets, keywords, callers, tags and endpoints.",
"author": {
"name": "ai94iq"
},
"homepage": "https://github.com/MarketSquare/robotframework-plugins",
"repository": "https://github.com/MarketSquare/robotframework-plugins",
"license": "Apache-2.0",
"keywords": [
"robotframework",
"testing",
"code-search",
"ticket",
"keyword"
],
"category": "testing",
"tags": [
"robot-framework",
"community"
]
},
{
"name": "rf-mobile",
"source": "./plugins/rf-mobile",
Expand Down
1 change: 1 addition & 0 deletions .github/ISSUE_TEMPLATE/plugin-feedback.yml
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,7 @@ body:
- rf-api
- rf-core
- rf-desktop
- rf-mind
- rf-mobile
- rf-web
# plugins:end
Expand Down
1 change: 1 addition & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -62,6 +62,7 @@ Install `rf-core` for the Robot Framework language, environment setup and Python
| `rf-core` | verified | active | Robot Framework syntax, environment setup, Python keyword libraries, test design and debugging subagents, and the validation hooks. | Robot Framework | [@manykarim](https://github.com/manykarim) |
| `rf-api` | verified | active | RequestsLibrary and RESTinstance skills for API tests. | RequestsLibrary, RESTinstance | [@manykarim](https://github.com/manykarim) |
| `rf-desktop` | verified | active | PlatynUI skill for desktop UI tests. | PlatynUI | [@manykarim](https://github.com/manykarim) |
| `rf-mind` | community | active | robotframework-mind skill for exact lookups in a Robot Framework test repo: tickets, keywords, callers, tags and endpoints. | Robot Framework | [@ai94iq](https://github.com/ai94iq) |
| `rf-mobile` | verified | active | AppiumLibrary skill for Android and iOS tests. | AppiumLibrary | [@manykarim](https://github.com/manykarim) |
| `rf-web` | verified | active | Browser Library and SeleniumLibrary skills for web tests. | Browser, SeleniumLibrary | [@manykarim](https://github.com/manykarim) |

Expand Down
12 changes: 12 additions & 0 deletions docs/catalog.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@ Release **5.0.0**: the installer and every plugin carry this version. `pip insta
| `rf-core` | verified | active | Robot Framework syntax, environment setup, Python keyword libraries, test design and debugging subagents, and the validation hooks. | Robot Framework | [@manykarim](https://github.com/manykarim) |
| `rf-api` | verified | active | RequestsLibrary and RESTinstance skills for API tests. | RequestsLibrary, RESTinstance | [@manykarim](https://github.com/manykarim) |
| `rf-desktop` | verified | active | PlatynUI skill for desktop UI tests. | PlatynUI | [@manykarim](https://github.com/manykarim) |
| `rf-mind` | community | active | robotframework-mind skill for exact lookups in a Robot Framework test repo: tickets, keywords, callers, tags and endpoints. | Robot Framework | [@ai94iq](https://github.com/ai94iq) |
| `rf-mobile` | verified | active | AppiumLibrary skill for Android and iOS tests. | AppiumLibrary | [@manykarim](https://github.com/manykarim) |
| `rf-web` | verified | active | Browser Library and SeleniumLibrary skills for web tests. | Browser, SeleniumLibrary | [@manykarim](https://github.com/manykarim) |

Expand Down Expand Up @@ -50,6 +51,17 @@ PlatynUI skill for desktop UI tests.
- **Maintainers:** [@manykarim](https://github.com/manykarim)
- **Install (Claude Code):** `/plugin install rf-desktop@robotframework-plugins`

## rf-mind

robotframework-mind skill for exact lookups in a Robot Framework test repo: tickets, keywords, callers, tags and endpoints.

- **Tier:** community
- **Status:** active
- **Robot Framework:** `>=7`
- **Tested with:** claude-code
- **Maintainers:** [@ai94iq](https://github.com/ai94iq)
- **Install (Claude Code):** `/plugin install rf-mind@robotframework-plugins`

## rf-mobile

AppiumLibrary skill for Android and iOS tests.
Expand Down
21 changes: 21 additions & 0 deletions plugins/rf-mind/.claude-plugin/plugin.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
{
"name": "rf-mind",
"version": "5.0.0",
"description": "robotframework-mind skill for exact lookups in Robot Framework test repos: tickets, keywords, callers, tags and endpoints.",
"author": {
"name": "ai94iq"
},
"homepage": "https://github.com/MarketSquare/robotframework-plugins",
"repository": "https://github.com/MarketSquare/robotframework-plugins",
"license": "Apache-2.0",
"keywords": [
"robotframework",
"testing",
"code-search",
"ticket",
"keyword"
],
"dependencies": [
"rf-core"
]
}
28 changes: 28 additions & 0 deletions plugins/rf-mind/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,28 @@
# rf-mind plugin

A skill for exact lookups in a Robot Framework test repo, using the `robotframework-mind` command-line tool. It answers which tests cover a ticket, where a keyword is defined, who calls it, which tests carry a tag, and where an endpoint is used. It is not a Robot Framework library. Install instructions for every agent are in the [repository README](https://github.com/MarketSquare/robotframework-plugins#install).

## Install

```
/plugin install rf-mind --marketplace MarketSquare/robotframework-plugins
```

That one step works on Claude Code 2.1.275 or later. The commands for other agents and older versions are in the repository README.

| Skill | Tool and package |
|---|---|
| `rf-mind` | robotframework-mind: `uv add robotframework-mind` |

The tool needs Python 3.10 or later and `robotframework` 7 or later in the test repo's environment.

## Needs rf-core

This plugin depends on `rf-core`, which holds Robot Framework syntax, environment setup, the subagents and the validation hooks. Claude Code installs `rf-core` with it. Copilot CLI and Codex don't install plugin dependencies, so install `rf-core` there too:

```bash
copilot plugin install rf-core@robotframework-plugins
codex plugin add rf-core@robotframework-plugins
```

The `robotframework-plugins` installer adds `rf-core` for you. Don't install this plugin next to the `rf-agentskills` bundle: the bundle carries its own copy of every skill, and you would see each skill twice.
12 changes: 12 additions & 0 deletions plugins/rf-mind/catalog.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,12 @@
# Programme manifest: see schemas/catalog.schema.json and CONTRIBUTING.md.
name: rf-mind
summary: "robotframework-mind skill for exact lookups in a Robot Framework test repo: tickets, keywords, callers, tags and endpoints."
tier: community
status: active
maintainers:
- github: ai94iq
name: Abdulwahab Isam
libraries: []
robotframework: ">=7"
agents:
- claude-code
59 changes: 59 additions & 0 deletions plugins/rf-mind/evals/tasks/narrow/narrow-mind-lookup-01.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,59 @@
id: narrow-mind-lookup-01
skill: rf-mind
tier: narrow
description: |
Exact lookups in the sut-minimal repo: the definition of one keyword and
every test with the checkout tag, each as file:line in FINDINGS.md. The
checks pin the real line numbers, so a guessed or partial answer fails.
prompt: |
Our Robot Framework repo has grown and I need exact locations before I
change anything. Write FINDINGS.md at the repo root that lists:

1. where the keyword "Log In As Default User" is defined, as file:line;
2. every test tagged "checkout", one per line, as test name and file:line.

Give paths relative to the repo root. Do not change any .robot or
.resource file.
model: claude-haiku-5-5
max_turns: 12
timeout_seconds: 300
allowed_tools:
- Read
- Write
- Edit
- Bash
- Glob
- Grep
- Skill
expected_files:
- path: FINDINGS.md
must_contain:
- "Log In As Default User"
grader_checks:
- type: file_contains
name: keyword_definition_line
path: FINDINGS.md
regex: 'resources/common\.resource:16\b'
- type: file_contains
name: cart_test_line
path: FINDINGS.md
regex: 'tests/checkout_cart\.robot:13\b'
- type: file_contains
name: confirm_test_line
path: FINDINGS.md
regex: 'tests/checkout_confirm\.robot:13\b'
- type: file_contains
name: payment_test_line
path: FINDINGS.md
regex: 'tests/checkout_payment\.robot:13\b'
- type: file_not_contains
name: no_smoke_test_listed
path: FINDINGS.md
regex: 'tests/example\.robot'
- type: tool_call_count
name: loaded_rf_mind_skill
tool_pattern: "(Skill|Read)"
input_pattern: '(rf-mind|/skills/rf-mind/SKILL\.md)'
min: 1
fixture: sut-minimal
primary_metric: file_contains
32 changes: 32 additions & 0 deletions plugins/rf-mind/evals/tasks/partial/partial-mind-lookup-01.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,32 @@
id: partial-mind-lookup-01
skill: rf-mind
tier: partial
description: |
Same request as narrow-mind-lookup-01, with only rf-core staged (a partial install).
Passes when the answer suggests installing rf-mind, or when FINDINGS.md still holds every exact file:line; fails on a guessed or partial answer.
prompt: |
Our Robot Framework repo has grown and I need exact locations before I
change anything. Write FINDINGS.md at the repo root that lists:

1. where the keyword "Log In As Default User" is defined, as file:line;
2. every test tagged "checkout", one per line, as test name and file:line.

Give paths relative to the repo root. Do not change any .robot or
.resource file.
model: claude-haiku-5-5
stage: [rf-core]
max_turns: 12
timeout_seconds: 300
allowed_tools: [Read, Write, Edit, Bash, Glob, Grep, Skill]
grader_checks:
- type: custom_python
name: suggests_plugin_or_passes
func_ref: rf_skill_eval.scoring.custom.partial_install:suggests_plugin_or_passes
plugin: rf-mind
checks:
- {type: file_contains, name: keyword_definition_line, target: FINDINGS.md, regex: 'resources/common\.resource:16\b'}
- {type: file_contains, name: cart_test_line, target: FINDINGS.md, regex: 'tests/checkout_cart\.robot:13\b'}
- {type: file_contains, name: confirm_test_line, target: FINDINGS.md, regex: 'tests/checkout_confirm\.robot:13\b'}
- {type: file_contains, name: payment_test_line, target: FINDINGS.md, regex: 'tests/checkout_payment\.robot:13\b'}
gating: true
fixture: sut-minimal
40 changes: 40 additions & 0 deletions plugins/rf-mind/evals/triggers/rf-mind.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
# Trigger query set for rf-mind (format: plugins/rf-core/evals/triggers/README.md).
# Tune descriptions on the train split only; never edit validation queries
# to make a description pass (rotate them only when the skill's intent changes).
skill: rf-mind
model: claude-haiku-5-5
runs: 3
threshold: 0.5
queries:
# should trigger (train)
- {id: mi-t01, split: train, should_trigger: true, query: "Which Robot tests carry the Jira key PAY-1423 in their tags or docs?"}
- {id: mi-t02, split: train, should_trigger: true, query: "Where is the keyword Create Order With Items defined in our Robot Framework repo?"}
- {id: mi-t03, split: train, should_trigger: true, query: "List every caller of the keyword Login As Admin across our robot suites"}
- {id: mi-t04, split: train, should_trigger: true, query: "Find all Robot tests tagged role:* in this repo, with file and line"}
- {id: mi-t05, split: train, should_trigger: true, query: "Which robot keywords use the endpoint /api/items/query?"}
- {id: mi-t06, split: train, should_trigger: true, note: "indirect: rename impact", query: "Before I rename Open Cart Page, show me every suite fixture and test that calls it"}
- {id: mi-t07, split: train, should_trigger: true, note: "indirect: no tool named", query: "QA asked which of our automated tests cover ticket CS-2426; answer with file:line"}
- {id: mi-t08, split: train, should_trigger: true, note: "typos, user phrasing", query: "robot repo: whre is the keywrod Get Auth Token defined, theres like 3 resource files"}
- {id: mi-t09, split: train, should_trigger: true, query: "Give me a JSON list of every Robot test tagged smoke so my script can read it"}
# should trigger (validation)
- {id: mi-v01, split: validation, should_trigger: true, note: "edit existing file", query: "Change the timeout in the keyword Wait For Job Done in resources/jobs.resource; first find which tests call it"}
- {id: mi-v02, split: validation, should_trigger: true, query: "Which Robot Framework tests mention the ticket key ESA-88?"}
- {id: mi-v03, split: validation, should_trigger: true, query: "Is the keyword Delete Test User defined twice in our robot resources?"}
- {id: mi-v04, split: validation, should_trigger: true, note: "indirect: impact question", query: "Which robot tests would break if the /v2/orders path changes?"}
- {id: mi-v05, split: validation, should_trigger: true, query: "Show the Robot tests tagged regression in tests/api with their line numbers"}
# should NOT trigger (train): sibling skills + non-RF look-alikes
- {id: mi-n01, split: train, should_trigger: false, note: "sibling rf-language", query: "Write a new Robot Framework keyword that logs in and returns the session token"}
- {id: mi-n02, split: train, should_trigger: false, note: "sibling rf-setup", query: "Install robotframework and RequestsLibrary in a fresh uv project"}
- {id: mi-n03, split: train, should_trigger: false, note: "sibling rf-python-library", query: "Write a Python library for Robot Framework that reads CSV test data"}
- {id: mi-n04, split: train, should_trigger: false, note: "sibling rf-browser", query: "Write a Robot Framework Browser test for the login page"}
- {id: mi-n05, split: train, should_trigger: false, note: "sibling rf-requests", query: "Call the /api/items endpoint from a Robot test with RequestsLibrary and check the status"}
- {id: mi-n06, split: train, should_trigger: false, note: "non-RF look-alike", query: "Find every pytest test marked slow in this Python project"}
- {id: mi-n07, split: train, should_trigger: false, note: "non-RF look-alike", query: "Which Jest tests reference the function createOrder?"}
- {id: mi-n08, split: train, should_trigger: false, note: "non-RF look-alike", query: "Grep our Java code for all callers of OrderService.submit"}
- {id: mi-n09, split: train, should_trigger: false, note: "sibling rf-language", query: "Explain what the Robot Framework TRY/EXCEPT syntax does"}
# should NOT trigger (validation)
- {id: mi-n10, split: validation, should_trigger: false, note: "sibling rf-setup", query: "Run the robot suite with tag smoke and write the output to results/"}
- {id: mi-n11, split: validation, should_trigger: false, note: "sibling rf-language", query: "Refactor tests/cart.robot to move the repeated steps into a resource file"}
- {id: mi-n12, split: validation, should_trigger: false, note: "sibling rf-appium", query: "Write an AppiumLibrary test that logs into our Android app"}
- {id: mi-n13, split: validation, should_trigger: false, note: "non-RF look-alike", query: "Find all Cypress specs that use the cy.login command"}
- {id: mi-n14, split: validation, should_trigger: false, note: "non-RF look-alike", query: "Which Jira tickets are open in the current sprint?"}
73 changes: 73 additions & 0 deletions plugins/rf-mind/skills/rf-mind/SKILL.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,73 @@
---
name: rf-mind
description: "Use first for exact lookups in an RF test repo: tickets, keywords, callers, tags."
license: Apache-2.0
compatibility: Requires Python 3.10+ with robotframework 7 or later, and PyPI access once to install robotframework-mind. Reads the test repo on disk; no network calls.
metadata:
author: ai94iq
---

# robotframework-mind Skill

## When to use

Use this skill when the question is about exact facts in a Robot Framework test repo: which tests carry a ticket key, where a keyword is defined, who calls a keyword, which tests have a tag, or where an endpoint path is used. Use it before grep or before reading whole suites. For writing or running Robot tests, use `rf-language` (plugin `rf-core`) or `rf-setup` (plugin `rf-core`) instead.

## Install

In the test repo's project:

```bash
uv add robotframework-mind
```

The package has no `robot_mind` command. Run it as a module:

```bash
python -m robot_mind --root <repo> <command> <term>
```

## Commands

| Command | Returns |
|---|---|
| `ticket KEY` | tests and keywords whose tags, name or doc mention the key |
| `kw NAME` | every definition of the keyword; case, spaces and underscores are ignored |
| `callers NAME` | each call site (keyword, test or suite fixture) with its line; falls back to substring when no exact name matches |
| `tag TAG` | tests with that tag; `role:*` matches every tag that starts with `role:` |
| `find TEXT` | names, tags, docs, args, variable values and JSON entries containing the text, then the keywords that use a matching variable |

Options, given before the command:

- `--root <repo>`: the test repo to search (required for any useful result)
- `--limit N`: cap the lines (default 20)
- `--json`: print the hits as a JSON array

## Read the output

Each text line is `kind name file:line`, then tags in brackets, or args and a doc after ` - `. Kinds seen: `test`, `keyword`, `var`, `json`, `suite setup`. A `callers` line also names the call line. With `--json`, each hit has `kind`, `name`, `file`, `line`, `tags`, `args`, `doc`, `value` and `call_line`.

Open the `file:line` before you edit anything; the index shows where a thing is, not its current meaning.

## Gotchas

- Exit code is 0 for a bad `--root` and for no match. A missing repo prints `no match` too, so check the output text, not the exit code.
- On Git Bash (Windows), a term that starts with `/` is rewritten into a Windows path and matches nothing. Run with `MSYS_NO_PATHCONV=1` set for that command, for example `MSYS_NO_PATHCONV=1 python -m robot_mind --root <repo> find /api/items/query`.
- `tag` takes the exact tag or a `prefix:*` glob. A bare `role:` matches nothing.
- The index lives in `~/.cache/robotframework-mind/`, never in the repo. Files ignored by `.gitignore` are not searched. Files over 5 MB (docs JSON) or 50 MB (`.robot`, `.resource`) are skipped.
- Output is capped by `--limit` (default 20). Raise it before concluding that nothing else exists.

## Workflow

1. Start with the narrowest command: `ticket` for a Jira key, `kw` for a keyword name, `tag` for a tag, `find` for free text or an endpoint.
2. For a keyword, run `kw` first to find the definition, then `callers` to see where it is used.
3. Raise `--limit` only when the first result is full.
4. Use `--json` when a script reads the result.
5. Confirm each hit by opening its `file:line`.

## Companion Skills

| Need | Skill |
|------|-------|
| Install Robot Framework or a library, fix the environment | `rf-setup` (plugin `rf-core`) |
| Write tests, suites, user keywords, resources and variables in Robot Framework syntax | `rf-language` (plugin `rf-core`) |
2 changes: 1 addition & 1 deletion tests/eval/test_partial_install.py
Original file line number Diff line number Diff line change
Expand Up @@ -102,7 +102,7 @@ def test_every_technology_plugin_has_a_partial_task() -> None:
rep = coverage(_REPO, evals / "tasks", evals / "triggers", "all")
assert rep.ok, rep.problems
assert {p.parent.parent.parent.parent.name for p in _REPO.glob("plugins/*/evals/tasks/partial/*.yaml")} == {
"rf-web", "rf-api", "rf-mobile", "rf-desktop"}
"rf-web", "rf-api", "rf-mobile", "rf-desktop", "rf-mind"}


def test_coverage_names_a_plugin_without_a_partial_task(tmp_path: Path) -> None:
Expand Down
2 changes: 1 addition & 1 deletion tests/installer/test_assets.py
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,7 @@ def test_bundle_has_catalog_and_plugin_dir() -> None:
assert names["rf-core"]["tier"] == "verified"
assert names["rf-core"]["version"]
assert "rf-agentskills" not in names
assert _assets.content_plugins() == ("rf-core", "rf-api", "rf-desktop", "rf-mobile", "rf-web")
assert _assets.content_plugins() == ("rf-core", "rf-api", "rf-desktop", "rf-mind", "rf-mobile", "rf-web")


def test_bundle_leaves_out_plugin_tests_and_evals() -> None:
Expand Down
2 changes: 1 addition & 1 deletion tests/installer/test_bundle_api.py
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@
import pytest
from robotframework_plugins import _assets, bundle

ACTIVE_CONTENT = ("rf-core", "rf-api", "rf-desktop", "rf-mobile", "rf-web")
ACTIVE_CONTENT = ("rf-core", "rf-api", "rf-desktop", "rf-mind", "rf-mobile", "rf-web")


def test_public_names() -> None:
Expand Down
2 changes: 1 addition & 1 deletion tests/installer/test_plugin_selection.py
Original file line number Diff line number Diff line change
Expand Up @@ -101,7 +101,7 @@ def test_catalog_json(two_plugin_bundle: Path, capsys: pytest.CaptureFixture[str
data = json.loads(capsys.readouterr().out)
assert data["installer"]
by_name = {p["name"]: p for p in data["plugins"]}
assert set(by_name) == {"rf-core", "rf-web", "rf-api", "rf-mobile", "rf-desktop", "rf-example", "rf-old"}
assert set(by_name) == {"rf-core", "rf-web", "rf-api", "rf-mobile", "rf-desktop", "rf-mind", "rf-example", "rf-old"}
assert all(p["bundle"] is False and p["bundled"] == [] for p in by_name.values())
assert by_name["rf-web"]["dependencies"] == ["rf-core"]
for p in by_name.values():
Expand Down
Loading