| name | standardize |
| description | Define or update a standard and apply it consistently across two scripts. |
Defines (or updates) a standard and applies it consistently across two scripts.
$ARGUMENTS
Process
Step 0: Gather standardization requirements
Gather the following information from the user-provided context:
- Which two scripts should be standardized (paths under
.agents/skills/).
- What specifically should be standardized (examples: step structure, decision-point structure, wording conventions, what belongs in Step 0, verification/commit conventions, shared checklist).
- Where the standard should live (the canonical home). Examples:
- A documentation page under
docs/docs/ (preferred for shared human+AI guidance)
- A command script under
.agents/skills/ that acts as the living spec for this class of scripts
- A code module/config if the “standard” is actually behavior (not prose)
Regardless, the standard has to live somewhere if the standard is to survive past this particular conversation. The standard has to be a long-lived artifact that can be explicitly referenced, so that when the standard changes, all applicable downstream artifcats also change with it.
- Any constraints about what must not change (behavior, wording, ordering, “don’t touch this user-edited section”, etc.).
If any information is unclear or missing, ask the user for details.
Step 1: Compare the scripts and identify what needs a standard
Read both scripts and identify:
- What pattern is currently inconsistent (structure, sequencing, terminology, decision points, verification/commit steps).
- What is genuinely shared vs. what is intentionally script-specific.
- Whether there is already an explicit standard for this (and where), or whether it needs to be written down.
Focus on standardizing decisions and structure, not on forcing identical wording.
Step 2: Define the standard (before changing scripts)
Propose the standard explicitly (the rules), and place it in the agreed canonical home.
This command is for defining standards. If a standard already exists but is incomplete/unclear, improve it as part of this step.
Iterate until the user approves the standard. Do not change the target scripts until the standard itself is agreed and written down.
Step 3: Apply the standardization
Update both scripts so they conform to the defined standard while preserving intended script-specific behavior.
Important: choose explicitly whether scripts should embody the standard, link to the standard (“see also”), or do a mix. All are valid:
- Embody: scripts follow the standard structurally without referencing it.
- Link: scripts include a short “see also” pointer to the canonical standard so the standard is discoverable at point-of-use.
- Mix: embody the structural rules and add a brief link for discoverability.
Step 4: Update documentation about the scripts (if applicable)
If the new/changed behavior impacts user-facing understanding of commands, update:
docs/docs/standards/workflow.md if the standardized commands are standard workflow commands (used by all mekara projects)
docs/docs/development/workflows.md if the standardized commands are mekara-specific
Step 5: Commit (only with user confirmation)
Summarize:
- What standard was defined/updated and where it lives
- What changed in each target script to conform
- Any intentionally script-specific deviations
Ask the user if they want to commit. If yes, use the committer agent to stage and commit all related changes (do not stage manually unless you are Codex).
Key Principles
- Define before applying: Do not modify the target scripts until the standard is written down and the user has approved it. Applying a half-baked standard wastes effort and creates inconsistency.
- Standards need a canonical home: The standard must be a durable artifact that can be explicitly referenced. Without a home, the standard only exists for the duration of the conversation.