validate-dod
Skill0 starsChanged 8 months agoMITView on GitHub
Builds, runs, and validates the implementation against the task's Definition of Done, then fills in the Execution Report.
In this repository11 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- code-reviewSkill
- cpp-opencl-devSkill
- design-creationSkill
- gradingSkill
- readme-creationSkill
- sync-designSkill
- sync-readmeSkill
- task-creationSkill
- technical-auditSkill
- ux-reviewSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Read the task file to extract: build directory, binary name, and every DoD checkbox.
- Very specific: line 14
Read the task's ## Relevant Files to locate the module root.
- Distinctive: line 24
Run each DoD item that requires execution.
- Very specific: line 39
Update the ## Execution Report section of the task file only.
- Very specific: line 43
Check each - [ ] DoD checkbox that passed → - [x]
- Very specific: line 44
Leave failed checkboxes as - [ ] and append a note explaining the failure
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
DoD Validation Skill
You are executing the Validation phase. Your job is to confirm the implementation works correctly at runtime and record the results. You MUST NOT modify any source files (.cpp, .cl, CMakeLists.txt).
Instructions
1. Read Context
- Read the task file to extract: build directory, binary name, and every DoD checkbox.
- Read the task's
## Relevant Filesto locate the module root.
2. Build
cdinto the module directory and run:cmake -B build && cmake --build build 2>&1- Capture full output. If the build fails, record the error in the Execution Report and stop — do not proceed to runtime checks.
3. Runtime Validation
Run each DoD item that requires execution. For each check:
| DoD pattern | Command / assertion |
| Binary runs without arguments | ./build/<binary> — exit code must be 0 |
output.bmp produced | ls -lh output.bmp after run |
| Timing table printed | grep for the required row labels in stdout |
| All kernel times non-zero | grep rows, assert values are not 0.000 |
--help flag works | ./build/<binary> --help — exit code 0, flags visible |
| Pixel verification passes silently | no [ERROR] lines in stdout |
| GPU env var respected | GPU=<vendor> ./build/<binary> — no crash |
Capture stdout and stderr for each run.
4. Fill Execution Report
Update the ## Execution Report section of the task file only. Use FileEdit. Set:
**Status:**→COMPLETE(all DoD pass) orFAILED(any DoD fails)**Session:**→ today's date (YYYY-MM-DD)### Validationcode block → trimmed stdout of the main run (≤ 40 lines)- Check each
- [ ]DoD checkbox that passed →- [x] - Leave failed checkboxes as
- [ ]and append a note explaining the failure
5. Communication Rules
- Terse. No filler.
- Report only: Status, which DoD items passed/failed, and the trimmed validation output.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.