원클릭으로
validate
Mark files as human-reviewed with specific validation types
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Mark files as human-reviewed with specific validation types
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | validate |
| description | Mark files as human-reviewed with specific validation types |
| argument-hint | <file-or-directory> <validation-type> [validation-type...] [--notes 'context'] |
| allowed-tools | Read Write StrReplace Shell Glob Grep |
| metadata | {"claude-tools":["Read","Write","Write","Bash","Glob","Grep"]} |
Parse $ARGUMENTS to extract:
read, fact-checked, tested, commands-verified, used-in-practice, sources-checked, voice-approved--notes flagExamples:
/validate docs/ai-engineering/the-shift.md read/validate docs/philosophy/ego-ai-and-the-zen-antidote.md read voice-approved/validate devops/ocp/troubleshooting/api-slowness-web-console/ read commands-verified --notes "Verified on OCP 4.14"/validate .agents/skills/backlog/SKILL.md read used-in-practiceIf no validation type is provided, ask: "What kind of validation? Options: read, fact-checked, tested, commands-verified, used-in-practice, sources-checked, voice-approved"
Special note on voice-approved: This type means the author has reviewed content that speaks in their voice — biographical claims, professional identity, personal opinions, experience statements. If a docs/ file contains biographical content and the user validates with read but not voice-approved, note: "This file contains biographical content. Consider also running /validate <path> voice-approved after reviewing those sections."
If the target is a directory, expand to all .md files in that directory (non-recursive by default; ask if the user wants recursive).
For each file:
---)review: block already exists within the frontmatterShow the user what will be updated:
## Validation Preview
Recording at: abc1234 (current HEAD)
### file-path.md
- Current status: [unreviewed / no metadata (legacy direction-reviewed) / direction-reviewed / reviewed]
- Adding: read (2026-04-18), commands-verified (2026-04-18)
- New status: reviewed
- Notes: "Verified on OCP 4.14"
### another-file.md
- Current status: reviewed (read: 2026-04-15, at: def5678)
- Adding: fact-checked (2026-04-18)
- Updating at: def5678 → abc1234
- Notes: none
Proceed? [y/n]
For each confirmed file:
First, capture the current short SHA: git rev-parse --short HEAD
If file has no frontmatter — add a new frontmatter block at the top:
---
review:
status: reviewed
read: 2026-04-18
at: abc1234
---
If file has frontmatter but no review: block — add review: to the existing frontmatter block.
If file already has a review: block — merge new validation types with existing ones. Do not overwrite existing dates; only add new types or update types explicitly being re-validated. Always update at to the current SHA — it marks the point-in-time of the most recent validation.
Set status: reviewed whenever any validation type is present.
For files in docs/ that have the standard AI disclosure footer:
If the footer says "has not been fully reviewed by the author", offer to update it:
"This file now has review metadata. Want me to update the disclosure footer to say it has been reviewed?"
If yes, replace:
After applying all changes, show:
## Validation Complete
Updated N files:
- file-path.md — reviewed (read, commands-verified)
- another-file.md — reviewed (read, fact-checked)
Disclosure footers updated: M files
<success_criteria>
at field set to current short SHA (git rev-parse --short HEAD)Review all pending changes against repo conventions before committing
Harvest and locally merge changes from a paude container session. Use when pulling agent work from paude, running paude harvest, merging session commits, importing submodule changes from a container, or fixing failed harvest/submodule fetch.
Start a paude container session for fire-and-forget agent work. Use when creating a paude session, delegating a task to paude, running work in an isolated container, setting up a paude worktree, or launching pi/claude/gemini in podman.
Write a task spec file for a paude container agent. Use when delegating work to paude, preparing a paude prompt-file, writing fire-and-forget agent tasks, or scoping submodule work for harvest.
Diagnose a running or stalled paude container session. Use when paude seems stuck, harvest returned empty, checking paude status, debugging 403 proxy blocks, or deciding wait vs harvest vs reset.
Create or update a project BRIEF.md — scaffolds planning structure for work that crosses the 5-file / multi-session threshold