sanad-implementation
Skill3 starsChanged 21 days agoMITView on GitHub
Internal implementation stage selected by Sanad after routing a substantive feature, behavior change, bug fix or implementation continuation. Never select for teaching, explanations, status or checking work alone.
In this repository11 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- braintrustSkill
- demoSkill
- explain-prSkill
- full-output-enforcementSkill
- gateSkill
- impeccableSkill
- sanadSkill
- shipSkill
- specSkill
- worktreeSkill
The file
Files
Sanad implementation stage
Reuse the task engine and role assignments from engines.md; do not ask again. Every Codex role remains xhigh.
Continue the implementation branch of Sanad from the current completed stage. Read the implementation loop (Sanad Steps 1–5) and execution.md before continuing; load engineering-bar.md at Step 3 and the other references only at the step that names them. Follow the saved route (self-routed or boss-selected) and its applicable steps, linked references, review requirements, gate and publishing approval. A review route skips the expert panel; a braintrust route resumes the saved routing boss after the specialists. Do not infer a panel from this effort-setting skill. Reuse the existing task state and valid evidence; do not restart routing or completed stages.
This skill only activates the implementation effort within Sanad. It does not create a second workflow, spawn an agent, broaden the task or authorize publishing. Resolve Sanad's relative references against its own directory.
The invoking request must require implementation. A request to learn how to implement something stays in Sanad's teaching branch. If only a question remains, answer it without starting implementation. On the next user turn, enter through Sanad again so it can select effort for that request.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.