QA review

Score the support conversation in "$ARGUMENTS" and say what would have made it better.

Compute with code. Perform every calculation (totals, balances, ratios, variances, projections) with a script — Bash, Python or a spreadsheet formula written to a file — and include the computation alongside the result. Never do arithmetic mentally.

Prerequisites

The conversation: pasted text, or a file path in $ARGUMENTS. When $ARGUMENTS or the request names a project, use that name as <project> whether or not its folder exists yet in .singularskills/support/ (create it when you write). With no project named: with exactly one folder there, use it; with several, ask which; with none at all, ask for a project name.

Read whichever of .singularskills/support/<project>/02-taxonomy.md, 04-macros.md (voice and ## Policies), 03-knowledge.md and 05-escalation.md exist; none is required.

Append this line to .singularskills/log.jsonl before you write the review (create the folder and file if missing): {"ts":"<ISO timestamp>","skill":"support-qa-review","workflow":"support","phase":null,"project":"<project>","event":"started","artifact":"qa/<date>-<slug>.md","note":""}, where <date> is today as YYYY-MM-DD and <slug> is two to four words naming the topic, never the customer.

Process

1. Summarise the conversation in two lines: what the customer needed and how it ended.

2. Score each dimension from 1 to 4 (1 poor, 2 needs work, 3 good, 4 excellent), each with the quoted evidence from the agent's messages:

   • Accuracy — was every statement correct, as far as the knowledge base and policies show?
   • Resolution — was the customer's need met, or the next step made clear?
   • Tone — did it follow the voice in 04-macros.md, if any, and treat the customer with respect?
   • Effort — how much did the customer have to do, repeat or wait for?
   • Policy — did the agent offer or promise only what the ## Policies section allows?
   • Escalation — was the ticket escalated when 05-escalation.md says it should have been, and not otherwise? When the conversation handles a failure, also score the outcome, the process and the treatment as references/service-recovery.md defines them.

3. Name the two changes that would most improve the conversation, each with a rewritten example sentence.

4. Protect personal data: quote only the agent's words the evidence needs, and replace any names, emails, phone numbers and account or order numbers with placeholders.

5. Write the review to .singularskills/support/<project>/qa/<date>-<slug>.md with sections Summary, Scores (a table: dimension, score, evidence), Changes and Strengths.

Output

The review written to .singularskills/support/<project>/qa/<date>-<slug>.md. Present the total and the lowest-scoring dimension, and the two changes.

After you have written the review, append the started line of Prerequisites to .singularskills/log.jsonl with "event":"finished": {"ts":"<ISO timestamp>","skill":"support-qa-review","workflow":"support","phase":null,"project":"<project>","event":"finished","artifact":"qa/<date>-<slug>.md","note":""}.

The review opens with this frontmatter (agent is the agent you run in, or unknown; this utility has no gate, so phase is null and gate is none):

---
skill: support-qa-review
workflow: support
phase: null
project: <project>
started: <ISO timestamp>
finished: <ISO timestamp>
agent: <the agent you run in, or "unknown">
gate: none
---

The review ends with ## Decisions and ## Open questions, in this shape:

## Decisions
- <what was chosen> — <why> — <what was rejected>

## Open questions
- <what the next phase or a reviewer must settle>