How do I install the Commercial Metrics Narrative skill?

Run npx skillmds@latest add mk-joshuajallen/commercial-metrics-narrative in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Commercial Metrics Narrative safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Commercial Metrics Narrative?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Commercial Metrics Narrative free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Commercial Metrics Narrative?

mk-joshuajallen (@mk-joshuajallen) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Commercial Metrics Narrative by mk-joshuajallen

Name: Commercial Metrics Narrative
Rating: 4.5 (4 reviews)
Author: mk-joshuajallen

Commercial Metrics Narrative

Regenerate the monthly commercial metrics Highs/Lows narrative for a run date. Use when someone asks to regenerate, write, refresh or redo the commercial metrics narrative, the country narrative, or narrative_country.json for a given run date or month. Triggers on "regenerate the narrative for run date X", "write the commercial metrics narrative", "redo the narrative for August".

File contents

Commercial metrics narrative

Thin wrapper. The work happens in the commercial-metrics-narrator subagent, which owns the rules, the reference files and the deterministic computation step.

This exists so the invocation people already use keeps working after the move to a subagent. Nobody calling it needs to know it changed.

What to do

Hand off to the subagent, passing through whatever the caller gave you.

Agent(
  subagent_type: "commercial-metrics-narrator",
  run_in_background: false,
  prompt: "Write the commercial metrics narrative for run date <YYYYMMDD>.
           <anything else the caller said, verbatim>"
)

Run it in the foreground. The subagent will pause partway through to ask the user for business context to overlay on the numbers, and that question has to reach the user rather than a background task queue.

What you handle before handing off

A missing run date. If the caller did not give one, look for the most recent output/<YYYYMMDD>/commercial_metrics_onepager_data and confirm it with them in one line before handing off. Do not hand off with an unresolved date.

A different phrasing of the same job. "Redo August", "the country narrative", "narrative_country.json" all mean this. Hand off.

What you do not handle

Do not compute anything. Do not read the CSVs. Do not draft bullets and pass them down. The whole point of the subagent is that the computation is deterministic and the writing happens with the reference files loaded.

After it returns

Relay the subagent's report to the user in full, in particular any unresolved flag and any check that failed or warned. Those are the reason this process exists in the shape it does, and they are useless if they stop at the subagent boundary.

---name: commercial-metrics-narrativedescription: Regenerate the monthly commercial metrics Highs/Lows narrative for a run date. Use when someone asks to regenerate, write, refresh or redo the commercial metrics narrative, the country narrative, or narrative_country.json for a given run date or month. Triggers on "regenerate the narrative for run date X", "write the commercial metrics narrative", "redo the narrative for August".---# Commercial metrics narrativeThin wrapper. The work happens in the `commercial-metrics-narrator` subagent, which owns the rules, the reference files and the deterministic computation step.This exists so the invocation people already use keeps working after the move to a subagent. Nobody calling it needs to know it changed.## What to doHand off to the subagent, passing through whatever the caller gave you.```Agent( subagent_type: "commercial-metrics-narrator", run_in_background: false, prompt: "Write the commercial metrics narrative for run date <YYYYMMDD>. <anything else the caller said, verbatim>")```Run it in the foreground. The subagent will pause partway through to ask the user for business context to overlay on the numbers, and that question has to reach the user rather than a background task queue.## What you handle before handing off**A missing run date.** If the caller did not give one, look for the most recent `output/<YYYYMMDD>/commercial_metrics_onepager_data` and confirm it with them in one line before handing off. Do not hand off with an unresolved date.**A different phrasing of the same job.** "Redo August", "the country narrative", "narrative_country.json" all mean this. Hand off.## What you do not handleDo not compute anything. Do not read the CSVs. Do not draft bullets and pass them down. The whole point of the subagent is that the computation is deterministic and the writing happens with the reference files loaded.## After it returnsRelay the subagent's report to the user in full, in particular any unresolved flag and any check that failed or warned. Those are the reason this process exists in the shape it does, and they are useless if they stop at the subagent boundary.

Frequently asked questions