OpenHarness End-to-End Evals
HKUDS/OpenHarness
Validates OpenHarness features by running real multi-turn agent loops with live LLM calls against an unfamiliar codebase, checking actual tool execution.
Generates test plans and test cases from local requirements for APP, Web, and server.
Install Codexqa Testcase Generator
$ npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install openqa-cn/codexqa codexqa-testcase-generator --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/openqa-cn/codexqa.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/codexqa-testcase-generator .claude/skills/codexqa-testcase-generator && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "codexqa-testcase-generator" agent skill from https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generator into .claude/skills/codexqa-testcase-generator/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "codexqa-testcase-generator", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generatorType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install openqa-cn/codexqa codexqa-testcase-generator --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/openqa-cn/codexqa.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/codexqa-testcase-generator .agents/skills/codexqa-testcase-generator && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "codexqa-testcase-generator" agent skill from https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generator into .agents/skills/codexqa-testcase-generator/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "codexqa-testcase-generator", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install openqa-cn/codexqa codexqa-testcase-generator --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/openqa-cn/codexqa.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/codexqa-testcase-generator .cursor/skills/codexqa-testcase-generator && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "codexqa-testcase-generator" agent skill from https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generator into .cursor/skills/codexqa-testcase-generator/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "codexqa-testcase-generator", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/openqa-cn/codexqa.git --path skills/codexqa-testcase-generator--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install openqa-cn/codexqa codexqa-testcase-generator --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/openqa-cn/codexqa.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/codexqa-testcase-generator .gemini/skills/codexqa-testcase-generator && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "codexqa-testcase-generator" agent skill from https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generator into .gemini/skills/codexqa-testcase-generator/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "codexqa-testcase-generator", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install openqa-cn/codexqa codexqa-testcase-generatorInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/openqa-cn/codexqa.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/codexqa-testcase-generator .github/skills/codexqa-testcase-generator && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "codexqa-testcase-generator" agent skill from https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generator into .github/skills/codexqa-testcase-generator/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "codexqa-testcase-generator", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install openqa-cn/codexqa codexqa-testcase-generator --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/openqa-cn/codexqa.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/codexqa-testcase-generator .opencode/skills/codexqa-testcase-generator && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "codexqa-testcase-generator" agent skill from https://github.com/openqa-cn/codexqa/tree/main/skills/codexqa-testcase-generator into .opencode/skills/codexqa-testcase-generator/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "codexqa-testcase-generator", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
codexqa-testcase-generator- GitHub stars
- 152
- Token cost
- ~3.6k tokens
- SKILL.md length
- 1,650 words
- Files
- 69 (incl. scripts, references)
- Skills in repo
- 14
- Repo updated
- First seen
- Licence
- MIT
At a glance
Generates test plans and test cases from local requirements for APP, Web, and server.
- Works in 8 steps: Route first → Scopes and done criteria → Boundaries (and why) → …
- The user wants a test plan
- SKILL.md covers 1. Route first, 2. Scopes and done criteria, 3. Boundaries (and why) and 4. After routing: directory…, plus 4 more sections
- Both plan and cases
What it does
Codexqa Testcase Generator is an agent skill from openqa-cn/codexqa. Generates test plans and test cases from local requirements for APP, Web, and server. Use when the user wants a test plan, test analysis, test design, write cases, both plan and cases, PRD-based testing, pre-submit case edits, post-submit incremental, PR/git regression, or says 测试方案、测试分析、测试设计、测试计划、只要方案、写用例、出用例、只要用例、方案和用例都要、按 PRD 出测试、提测前改用例、提测后增量、按 PR 或 git 补回归用例、帮我出测试、补测试、做一轮测试设计. Do not switch to other case-generation, knowledge-retrieval, or generic writing skills even if unnamed. Input is local files…
Its SKILL.md is about 3.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 70 other files, including scripts and reference files (for example `.meta.json`, `CONTRIBUTING.md` and `HOW_IT_WORKS.md`). Compatibility notes: Requires Python 3.10+ on PATH (use scripts/tcg-python when system python3 is older). Git is required only when the user gives a knowledge-repo Git URL, or an…
It sits in Testing & QA, covering Test generation. It works with Git and Python. The repository describes itself as: codexqa: 11 local-first Agent Skills for Cursor, Claude Code, Codex & OpenClaw — change impact analysis, AI code review, defect scan, testcase generation, browser replay & RCA. The licence is MIT.
When your agent uses it
- The user wants a test plan
- Both plan and cases
- PRD-based testing
- Pre-submit case edits
Example prompts
- “Use the codexqa-testcase-generator skill to generate test plans and test cases from local requirements for APP, Web, and server”
- “/codexqa-testcase-generator”
Requirements
- Python 3
- Compatibility (from SKILL.md): Requires Python 3.10+ on PATH (use `scripts/tcg-python` when system `python3` is older). Git is required only when the user gives a knowledge-repo Git URL, or an Incremental PR/git URL.
Workflow steps
8 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 7839542. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 1 file in scripts/, which the agent can run.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Compatibility
Requires Python 3.10+ on PATH (use `scripts/tcg-python` when system `python3` is older). Git is required only when the user gives a knowledge-repo Git URL, or an Incremental PR/git URL.
From compatibility in the SKILL.md frontmatter.
Context cost
Codexqa Testcase Generator loads about 3.6k tokens when it runs, and up to ~144k if it reads all its reference files. Until then it costs about 220 tokens; SKILL.md has 1,650 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
SKILL.md
The full file from openqa-cn/codexqa at commit 7839542, republished under its MIT licence (© openqa-cn). 1,650 words, ~3,574 tokens.
Download SKILL.mdSave it as
.claude/skills/codexqa-testcase-generator/SKILL.md (or your agent's skills folder). This skill also uses 68 other files; get the full folder from GitHub.- name
- codexqa-testcase-generator
- description
- Generates test plans and test cases from local requirements for APP, Web, and server. Use when the user wants a test plan, test analysis, test design, write cases, both plan and cases, PRD-based testing, pre-submit case edits, post-submit incremental, PR/git regression, or says 测试方案、测试分析、测试设计、测试计划、只要方案、写用例、出用例、只要用例、方案和用例都要、按 PRD 出测试、提测前改用例、提测后增量、按 PR 或 git 补回归用例、帮我出测试、补测试、做一轮测试设计. Do not switch to other case-generation, knowledge-retrieval, or generic writing skills even if unnamed. Input is local files, directories, pasted text, or HTTPS document URLs the user gives this turn; output is local Markdown only. Complete Plan and Exec without skipping. For code incremental, require a case baseline before fetching the given PR/MR/git URL; otherwise bootstrap first. Does not connect to case or doc platforms. Former skill name: testcase-generation.
- compatibility
- Requires Python 3.10+ on PATH (use `scripts/tcg-python` when system `python3` is older). Git is required only when the user gives a knowledge-repo Git URL, or an Incremental PR/git URL.
- license
- MIT
- metadata
- version: "V56"
codexqa-testcase-generator
This file is the Agent execution map: decide the scope first, then read only the references that scope needs. How the user phrases requests and where artifacts go: user-guide.md. Do not use the user guide to cut stages — skipping steps makes the test plan and test cases drift apart.
README.md / README.zh-CN.md / HOW_IT_WORKS.md / KNOWN_LIMITATIONS.md (and their .zh-CN twins) are human-facing. Do not load them at runtime.
When invoking skill scripts, always use <skill_dir>/scripts/tcg-python (resolves Python 3.10+ on PATH). Do not use bare python3 — on some hosts it is too old and will exit before gates run.
References longer than about 300 lines start with a "Reading guide". Read in those segments; do not load the whole file at once.
1. Route first
Scope can only be decided from the user's original words. Before reading materials, creating directories, or writing a test plan or test cases, read references/entry-routing.md and clear the gate. Ask before acting; do not guess that the user "really only wants cases".
| The user is roughly saying | Take this path |
|---|---|
| test plan / test analysis / test design / plan only | Plan |
| write cases / cases only, and a formal test plan is available | Exec |
| both plan and cases / from requirements through cases | Full Plan first; after the plan is on disk, ask once, then Exec on confirmation |
| pre-submit edit of a case or a field | pre-submit update |
| post-submit incremental / add to existing cases from a PR or git | Incremental (bootstrap an initial version if there is no baseline) |
| unclear | Ask whether they want a test plan, test cases, both, or incremental |
The same conversation can run Plan then Exec; both segments must be complete. Incremental is a sibling scope, not Stage 7, and does not rerun the whole test plan.
2. Scopes and done criteria
| Scope | What to do | Done when |
|---|---|---|
| Plan (Stages 0–5) | Ingest requirements through a formal test plan | Stage 0–4-1 artifacts + last check_run_gate.py --gate stage5 stdout ok: true + last close_stage.py --stage 5 stdout ok: true. The file alone is not done. |
| Exec (Stage 6) | Write local test cases from the formal test plan | {run_dir}/testcase/initialcase/ and dual-write cases/ + {run_dir}/testdesign/testcase_generation_report.html + last check_run_gate.py --gate stage6 stdout ok: true + last close_stage.py --stage 6 stdout ok: true |
| pre-submit update | Edit existing cases from the conversation; do not redo the test plan | {run_dir}/testcase/initialcase/ |
| Incremental | Diff-enhance or directly update on an existing baseline | {run_dir}/testcase/cases/ has been read back; the process area has decisions and a report |
3. Boundaries (and why)
These boundaries exist to avoid dropping a half-finished artifact and later patching it, and to keep the test plan, test cases, and incremental work from polluting each other.
- Plan does not write test cases; Exec does not redo the test plan. Cases must be anchored to the formal test plan already on disk; writing cases during the plan stage means they drift when the design later changes.
- Do not cut steps. "Just the result", "skip init", and "no knowledge pack" are not reasons to skip routing, audit, coverage, or persist. What is missing is facts, not stages.
- Exec accepts only Stage 5's
{run_dir}/testdesign/test_design.md. If the file is missing, unreadable, or has no final client-type decision, complete Plan first, then ask whether to write cases. Do not guess, and do not substitute a simplified case set. - Incremental does not change the formal test plan. Incremental maintains existing cases. If the user also wants a new test plan, run full Plan first, then decide Exec or incremental from confirmation.
- If information is missing, clarify or mark pending clarification / TBD. Fall back to the matching stage. Do not invent business facts. Do not call an external knowledge-retrieval Skill.
- Write local Markdown only.
run_diris created or reused by this Skill from the conversation. Do not connect to a case platform or a doc platform. - After the current scope is on disk and work receipts are written, ask about the next step.
4. After routing: directory and knowledge
Read references/shared-rules.md and references/run-workspace.md, then confirm run_dir (userConfig.runDir → user-specified → default $HOME/codexqa-testdata-generator/runs/{runid}). userConfig.runDir is the same path as the working directory; first write or a missing-field fill uses this default. After routing, the first script is close_stage.py --init, not writing the plan. Then read {run_dir}/testcase/testdocs/run-status.json and continue from currentStage. Do not jump to test_design.md. User-given http(s):// document URLs this turn are Stage 0 input: persist the fetched body under testcase/testdocs/; ingest_local_docs.py is optional.
Knowledge follows only references/knowledge-adapter.md:
Requirements / technical design / prior reports → built-in norms in this repo → a local knowledge directory or an explicitly given Git URL the user already provided (optional)
When the user actively pastes a knowledge-repo URL, shallow-clone it, then read the local index.md. An empty index is valid; continue with built-in norms. Write a work receipt for every actual action.
| Scope | Knowledge-read limit |
|---|---|
| Stage 0 | May bootstrap a local directory or user Git URL; empty index is valid. No business close reading |
| Stages 1–4 | Facts come from requirements / technical design; close-read only if a local pack exists |
| Stage 4-1 | Fill only gaps that prior stages did not cover and that the inputs can verify |
| Stage 5 | Integrate prior reports only; no new retrieval, no new scenarios |
| Stage 6 | Use the formal test plan and the three-end templates first; read the pack only if test-data setup / format gaps remain and a local pack already exists |
| Incremental | Use the formal test plan, the frozen baseline, and Chapter 3 of the three-end templates first; read the pack only if compatibility / regression / test-data setup gaps remain and a local pack already exists |
Show full SKILL.md (733 more words)Show less
5. Plan (test plan)
After the gate, choose a mode. Automatic: do not wait for the user between Stages 0–5. Manual: persist, summarize, and continue after confirmation. Neither mode may omit a stage or merge persist. Each stage must write its artifacts, run close_stage.py --stage N, then read only the next row’s files plus prior artifacts.
When entering a stage, read only that row's files. Each stage file opens with Minimum persist; the gate checks those headings.
| Stage | Read | Goal | Main artifacts |
|---|---|---|---|
| 0 | s00-req-ingest.md | Persist local or user-given HTTPS materials and validate | testcase/testdocs/, userConfig.json, run-status.json, Stage 0 receipts |
| 1 | s01-req-analysis.md | Identify terms, flows, rules, functions, and test objects from requirements | Stage 1 report, stage1-*/index.json |
| 2 | s02-impact-risk.md | Impact and risk from the technical design | Stage 2 report, stage2-*/index.json |
| 3 | s03-coverage-judge.md | Local coverage judgment; no remote space; treat all as new | Stage 3 report, stage3-*/index.json |
| 4 | s04-object-design.md, scene-fusion.md | Design scenarios with built-in models | Stage 4 report, stage4-*/index.json |
| 4-1 | s04a-design-audit.md | Three audit rounds; edit the Stage 4 report in place | stage4-1-audit-receipt.md + in-place Stage 4 report |
| 5 | s05-plan-compose.md, plan-md-template.md, plan-export-notes.md | Integrate prior conclusions only | testdesign/test_design.md, changelog |
After each Plan/Exec persist, run:
<skill_dir>/scripts/tcg-python <skill_dir>/scripts/close_stage.py --run-dir <userConfig.runDir> --stage <N>Before Stage 5 writes the plan, run --gate stage5. After Stage 6 dual-writes initialcase/ and cases/, run --gate stage6 (missing cases/ fails).
<skill_dir>/scripts/tcg-python <skill_dir>/scripts/check_run_gate.py --run-dir <userConfig.runDir> --gate stage5
<skill_dir>/scripts/tcg-python <skill_dir>/scripts/check_run_gate.py --run-dir <userConfig.runDir> --gate stage6Redirect stdout only (see shared-rules). If ok is false, do not write test_design.md, and do not claim Exec done. Resume from run-status.json currentStage or the first item in failures. The gate rejects heading-only reports, stub index.json, a missing or empty stage4-1-audit-receipt.md, empty Loop bodies, a plan missing Minimum persist headings or empty plan chapters, a plan whose S-xx set does not match Stage 4, and Stage 6 without cases/ dual-write. Stage 5 gate requires currentStage 5 and 4-1 in completed. Before saying Plan is done, this turn must have run --gate stage5 and you must cite that stdout ok: true. Before saying Exec is done, this turn must have run --gate stage6 and you must cite that stdout ok: true. Without that JSON, say draft.
After composing the plan, run <skill_dir>/scripts/tcg-python <skill_dir>/scripts/close_stage.py --run-dir <userConfig.runDir> --stage 5 (sets planPersistedAt, keeps execAllowed=false). Ask whether to generate test cases, then stop this turn. Do not run --allow-exec, do not read s06, and do not write cases in the same turn that first persisted test_design.md. A later user message that confirms Exec is required before --allow-exec.
6. Exec (test cases)
Start Exec only after a later user message that confirms Exec. Then run <skill_dir>/scripts/tcg-python <skill_dir>/scripts/close_stage.py --run-dir <userConfig.runDir> --allow-exec. If execAllowed is still false, do not write cases. Then read references/s06-case-write.md. Accept only the Stage 5 formal test plan. After dual-write to initialcase/ and cases/, run check_run_gate.py --gate stage6 via tcg-python, then close_stage.py --stage 6 via tcg-python. Without stage6 stdout ok: true, say draft.
Finish in order: client-type routing → necessary knowledge close reading → decompose and write from templates → coverage check → local persist (dual-write) → aggregated HTML report. Three-end writing follows Stage 6 and the matching case-tpl-*.md. After dual-write, run <skill_dir>/scripts/tcg-python <skill_dir>/scripts/generate_case_report.py --run-dir <userConfig.runDir> and cite the stdout out path ({run_dir}/testdesign/testcase_generation_report.html). The report aggregates Web / Server / APP case sets (missing ends show empty-state). Do not upload to a case platform. Do not write remote docs.
7. Pre-submit update
When the user wants to edit existing cases, and it is not a plan redo and not post-submit incremental, read references/presubmit-case-edit.md. Write only {run_dir}/testcase/initialcase/. Do not merge with Incremental.
8. Incremental (post-submit incremental)
Before starting, read references/postsubmit-enhance.md. Diff-enhance or directly update only on an existing valid baseline. If the target end has no baseline, bootstrap an initial version per references/postsubmit-bootstrap.md, then continue incremental.
Finish in order: identity confirm → execution workspace → freeze Chapter 3 of the three-end templates → archive inputs → change / intent analysis → per-end decisions → environment review (if any) → container placement read-back → publish. Details follow references/incremental/ and scripts/incremental/.
Valid baseline is defined in references/incremental/input-spec.md: user-explicit cases, or that end's testcase/cases/; read initialcase/ only when that cases/ path does not exist. If cases/ exists but is empty or corrupt, do not fall back.
When the user this turn explicitly gives a PR / MR, a Code-platform /code/repo-detail/.../pr/{n}, or a git repo URL, and routing is DIFF_ENHANCEMENT: confirm a valid baseline first. Do not call pull_remote_code.py without a baseline. With a baseline, fetch locally and archive per references/incremental/code-fetch-spec.md.
When explaining failures to the user, follow the script error.code:
PR_PERMISSION_DENIED: permission deniedPR_HEAD_REF_MISSING: ask for a local repo, branch name, or head SHA- other: report that error code
© openqa-cn, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 68 other files (scripts, references) in skills/codexqa-testcase-generator of openqa-cn/codexqa.
- SKILL.md
- .gitignore
- .meta.json
- CONTRIBUTING.md
- HOW_IT_WORKS.md
- HOW_IT_WORKS.zh-CN.md
- KNOWN_LIMITATIONS.md
- KNOWN_LIMITATIONS.zh-CN.md
- LICENSE
- README.md
- README.zh-CN.md
- references/case-export-notes.md
- references/case-tpl-app.md
- references/case-tpl-server.md
- references/case-tpl-web.md
- references/entry-routing.md
- references/incremental/artifact-spec.md
- references/incremental/case-standard-format.md
- references/incremental/code-fetch-spec.md
- … and 50 more
Open the folder on GitHubat commit 7839542
Compare with similar skills
Codexqa Testcase Generator next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Codexqa Testcase Generator this skillopenqa-cn/codexqa | 152 | — | ~3.6k | Automated safety check: Pass | MIT | |
| OpenHarness End-to-End EvalsHKUDS/OpenHarness | 16k | 1 repos | ~2.1k | Automated safety check: Notes | MIT | |
| Adk Verify Snippetsgoogle/adk-python | 22k | — | ~1.4k | Automated safety check: Pass | Apache-2.0 | |
| Update Megatron Golden ValuesNVIDIA/Megatron-LM | 18k | — | ~2.8k | Automated safety check: Pass | Custom licence | |
| Hermetic Python Unit TestsdimensionalOS/dimos | 4.6k | — | ~1.4k | Automated safety check: Pass | Custom licence | |
| Find Untested Sourcesdotnet/skills | 5.6k | 1 repos | ~3.3k | Automated safety check: Pass | MIT |
Similar skills
- Official
Adk Verify Snippets
google/adk-python
Checks that every Python code block in a Markdown file actually compiles and runs, by extracting each block to a temporary file, executing it in an isolated subprocess, and writing a pass/fail…
- Official
Update Megatron Golden Values
NVIDIA/Megatron-LM
Refreshes stored golden values from a GitHub Actions run, reports signed percentage changes per model, and writes a summary ready for a pull request description.
Hermetic Python Unit Tests
dimensionalOS/dimos
Rules for writing, fixing and reviewing pytest unit tests that are hermetic: behavior-focused, deterministic, isolated and cheap to run.
- Official
Find Untested Sources
dotnet/skills
Statically pairs source files with test files to list code that no test references, using Roslyn for C# or tree-sitter for many languages, with no build.
Test Coverage Review
areed1192/finance-news-aggregator
Audit, plan, write, and verify unit tests for Python projects using pytest.
More from openqa-cn/codexqa
All 14 skills in this repoCodexqa Code Wiki
openqa-cn/codexqa
Builds a local architecture wiki for a repository from the CodexQA symbol graph (no model needed): modules, who calls whom and how often, reading paths, and one self-contained HTML page.
Codexqa Rootcause Analyzer
openqa-cn/codexqa
Diagnoses exception root causes from stack traces, logs, call-chain dumps, and debug output using the CodexQA CLI for structured repo analysis.
Codexqa Testdata Generator
openqa-cn/codexqa
Constructs test data against real backends and writes it back into test cases as executable preconditions.
Codexqa Skill Router
openqa-cn/codexqa
Auto-routes a user request to the matching codexqa skill, then ensures that skill is on disk and follows its SKILL.md.
Catalog
openqa-cn/codexqa
Constructs catalog products (standard or limited), catalog orders, and account-credit enrollment, including product-then-credit scenes.
Codexqa Change Analysis
openqa-cn/codexqa
Change-impact analysis for one diff: use CodexQA diff indexing (index --diff-base tags nodes and files add / change / delete, then change-groups / symbol-diff / file-base) to write one HTML report…
Related
Categories
Questions about Codexqa Testcase Generator
What does Codexqa Testcase Generator do?
Generates test plans and test cases from local requirements for APP, Web, and server. Codexqa Testcase Generator is an agent skill from openqa-cn/codexqa. Generates test plans and test cases from local requirements for APP, Web, and server.
When should I use Codexqa Testcase Generator?
Codexqa Testcase Generator fits situations like: the user wants a test plan; both plan and cases; PRD-based testing; pre-submit case edits.
How do I install Codexqa Testcase Generator in Claude Code?
Run `npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a claude-code`. Or copy the skill folder (skills/codexqa-testcase-generator in openqa-cn/codexqa) into .claude/skills/codexqa-testcase-generator in your project. Claude Code loads it when a task matches its description.
How do I install Codexqa Testcase Generator in Codex?
Run `npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a codex`. Or copy the skill folder (skills/codexqa-testcase-generator in openqa-cn/codexqa) into .agents/skills/codexqa-testcase-generator in your project. Codex loads it when a task matches its description.
Can I use Codexqa Testcase Generator in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add openqa-cn/codexqa --skill codexqa-testcase-generator -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/codexqa-testcase-generator, .gemini/skills/codexqa-testcase-generator, .github/skills/codexqa-testcase-generator and .opencode/skills/codexqa-testcase-generator in your project.
What does Codexqa Testcase Generator need to run?
SKILL.md names no scripts, command-line tools or credentials: Codexqa Testcase Generator is instructions for the agent only. Our summary lists: Python 3. Compatibility (from SKILL.md): Requires Python 3.10+ on PATH (use `scripts/tcg-python` when system `python3` is older). Git is required only when the user gives a knowledge-repo Git URL, or an Incremental PR/git URL..
Does Codexqa Testcase Generator access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Codexqa Testcase Generator safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
What licence does Codexqa Testcase Generator use?
Codexqa Testcase Generator is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Codexqa Testcase Generator use?
About 3.6k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 140k tokens, read only when the agent opens those files.
What are the alternatives to Codexqa Testcase Generator?
Skills that share tags, products or a category with Codexqa Testcase Generator: OpenHarness End-to-End Evals (HKUDS/OpenHarness, 16k stars), Adk Verify Snippets (google/adk-python, 22k stars), Update Megatron Golden Values (NVIDIA/Megatron-LM, 18k stars) and Hermetic Python Unit Tests (dimensionalOS/dimos, 4.6k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Codexqa Testcase Generator?
openqa-cn (a GitHub organization) maintains it in openqa-cn/codexqa, which has 152 GitHub stars. The repository holds 14 skills in this directory. The repository was last updated on October 3, 2026.
Source: openqa-cn/codexqa on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.