misc-helper

name	misc-helper
description	Catch-all micro-skill for quick utility tasks that take under 5 minutes and don't fit any domain skill. Triggers: 'convert this file', 'reformat this data', 'quick calculation', 'transform this text', 'one-liner to do X', 'help me with this small thing'. Do NOT use when the task requires domain expertise (use the domain skill), involves multi-step workflows, touches production code, or would take more than 5 minutes.
license	Apache-2.0
compatibility	{"clients":["openai-codex","gemini-cli","opencode","github-copilot"]}
metadata	{"owner":"codex","domain":"misc-helper","maturity":"draft","risk":"low","tags":["utility","catch-all","quick-task","format-conversion","text-transform"]}

Purpose

A genuine catch-all micro-skill for one-off utility tasks that are too small to warrant a dedicated skill. Handles file format conversions, quick calculations, text transformations, data reformatting, one-liner scripts, and similar ephemeral work.

When to use this skill

Use this skill when:

• the task is a one-off utility job: format conversion (JSON↔YAML↔TOML↔CSV), text munging (regex replace, case conversion, encoding changes), quick math/date calculations, or data reshaping
• the task will take under 5 minutes of agent effort end-to-end
• no existing domain skill covers the task—you have checked the skill library first
• the user asks for a "quick script", "one-liner", or "just convert/reformat this"

Do not use this skill when

• the task requires domain expertise—use the domain skill instead (e.g., bigquery-skill for SQL optimization, fastapi-patterns for API design, tauri-solidjs for desktop app scaffolding)
• the task involves multi-step workflows with branching logic or significant decision-making
• the output will be committed as production code that needs tests, reviews, or maintenance
• the task would take more than 5 minutes—escalate to an appropriate domain skill or break it into tickets
• the task is a factual lookup with no transformation (just answer the question directly)

Operating procedure

1. Classify the task. Determine the category: format conversion, text transformation, quick calculation, data reshaping, or one-liner script.
2. Check for a better skill. Scan active skills to confirm no domain skill covers this task. If one does, redirect immediately.
3. Execute the minimal viable solution. Use the simplest tool for the job:
   • Format conversion → jq, yq, python -c, csvtool, or inline code
   • Text transformation → sed, awk, tr, or a short Python/Node snippet
   • Quick calculation → bc, python -c, or node -e
   • Data reshaping → jq, pandas one-liner, or shell pipes
4. Validate the output. Spot-check the result: correct format, no data loss, expected row/field count.
5. Return the result with a one-line description of what was done.

Decision rules

• 5-minute rule: if the task will take more than 5 minutes, stop and recommend the correct domain skill or suggest creating a ticket.
• Complexity escalation: if the task grows beyond the original ask (e.g., "also validate the schema" or "handle edge cases"), pause and reassess whether a domain skill should take over.
• No persistence: misc-helper results are ephemeral. If the user wants the transformation to be repeatable, recommend encoding it as a script or a new skill.
• Tool selection: prefer standard CLI tools (jq, sed, awk) over writing scripts. Prefer scripts over installing new dependencies.
• Idempotency: when transforming files in-place, always confirm with the user or create a backup first.

Scope boundaries

In scope:

• File format conversions (JSON↔YAML↔TOML↔CSV↔XML)
• Text transformations (regex, case, encoding, line ending normalization)
• Quick calculations (date math, unit conversion, base conversion)
• Data reshaping (flatten nested JSON, pivot CSV columns, extract fields)
• One-liner scripts (shell pipes, awk programs, jq filters)
• Quick lookups that require minimal synthesis (checking a file's encoding, counting lines, sampling data)

Out of scope:

• Anything requiring domain expertise (database optimization, API design, security review)
• Multi-step workflows with branching logic
• Production code that needs tests or maintenance
• Tasks that would benefit from a dedicated skill being created

Output requirements

Return exactly:

1. Result — the transformed data, calculation answer, or generated one-liner
2. What was done — a single sentence describing the transformation applied (e.g., "Converted config.yaml to JSON using yq and pretty-printed with jq")

Do not add commentary, recommendations, or follow-up suggestions unless the task revealed a problem.

Anti-patterns

• Skill avoidance: using misc-helper repeatedly for tasks in the same domain instead of learning or creating the proper domain skill. If you use misc-helper for the same category 3+ times, create a dedicated skill.
• Scope creep: letting a "quick conversion" grow into a multi-step data pipeline. Stop and redirect.
• Over-engineering: writing a 50-line script for a task jq '.[] | .name' would handle.
• Silent data loss: converting formats without validating that all fields survived the transformation.
• Dependency installation: installing new packages for a one-off task when a built-in tool would suffice.

Related skills

• bigquery-skill — for any data query or analytics task
• fastapi-patterns — for any Python API task
• tauri-solidjs — for any desktop application task

Failure handling

• If the task is ambiguous, classify it and state your interpretation before executing.
• If a format conversion loses data (e.g., YAML comments stripped in JSON conversion), warn the user and offer alternatives.
• If the task exceeds 5 minutes or grows in scope, stop and recommend the appropriate domain skill or suggest creating a ticket.
• If no CLI tool handles the format, fall back to a short Python/Node script and note the dependency.