ailtir-feedback

name	ailtir_feedback
description	Captures lightweight anonymous user feedback for an Ailtir workflow: first a 1-10 rating, then the reason, then three structured AskUserQuestion follow-ups derived from the current session.
allowed-tools	["mcp__plugin_ailtir-cowork-plugin_ailtir__plugin_report_usage","mcp__plugin_ailtir-cowork-plugin_ailtir__plugin_feedback"]

Ailtir Feedback

Usage Reporting

Before doing workflow-specific work, read the stable anonymous UUID from ~/Ailtir-Tendering/install_id. If the file is missing, create its parent directory, generate a UUID v4, and write only that UUID to the file. Re-read the file and call the plugin_report_usage tool from the bundled ailtir MCP server with these arguments:

• skill_name: ailtir_feedback
• plugin_version: 2.15.4
• installation_id: the UUID read from ~/Ailtir-Tendering/install_id

If the identifier cannot be read or created, leave that failure visible and continue without reporting. If reporting returns failed, leave the failure visible and continue the workflow.

You collect anonymous feedback without turning it into a long survey. The goal is to understand why the last Ailtir workflow was or was not useful.

Privacy Guardrail

Before asking for free text, say:

Avoid client names, prices, credentials, or tender-sensitive details. A short general reason is enough.

Do not ask for project names, client names, tender values, contact details, credentials, or commercially sensitive facts.

Step 1 - Ask For Rating First

Ask the user to rate the experience from 1 to 10.

Use this wording:

How useful was this Ailtir workflow from 1 to 10?

If the user gives anything outside 1-10, ask once for a valid number. If they decline or say skip, stop and do not report feedback.

Step 2 - Ask For The Reason Second

After the rating is known, ask:

What is the main reason for that rating? One sentence is enough.

If the user declines, continue with an empty reason. Do not pressure them for free text.

Step 3 - Analyze The Session

Review the current conversation and identify the workflow being rated. Prefer the most recent Ailtir skill. If that is unclear, use ailtir_feedback as the workflow name and skill as the workflow kind.

Analyze the rating, the reason, and the session context. Prepare exactly three follow-up questions that help explain the rating. Each question must:

• Be specific to what happened in the session.
• Have exactly three possible answers.
• Avoid asking for sensitive tender details.
• Be answerable in one click.

Good question themes include:

• whether the output was usable as-is, needed light edits, or needed major rework
• whether the workflow understood the tender context, partly understood it, or missed key context
• whether the biggest issue was accuracy, missing detail, speed, formatting, or integration
• whether the workflow saved time, broke even, or cost time

Step 4 - Use The Claude AskUserQuestion Tool

Use Claude Code's AskUserQuestion tool for the three follow-up questions when it is available. Make a single AskUserQuestion call containing exactly three questions. Each question must have exactly three options. Do not use multi-select.

Use option labels that are short and stable enough to report, for example:

• usable_as_is
• needed_light_edits
• needed_major_rework

Each option should include a short human-readable description.

If AskUserQuestion is unavailable in the current host, ask the same three questions in plain text with three numbered options each.

Step 5 - Submit Feedback

Call the plugin_feedback tool from the bundled ailtir MCP server with:

• rating: the 1-10 rating
• plugin_version: 2.15.4
• installation_id: the UUID read from ~/Ailtir-Tendering/install_id
• reason: the user's reason, or an empty string
• workflow_name: the identified Ailtir skill, plugin, or session
• workflow_kind: skill, plugin, or session
• followup_answers: the three selected answers keyed by stable question labels

Do not write feedback to Daily/feedback.md or another local file. If submission returns failed, leave the failure visible but do not retry or interrupt the workflow. Briefly thank the user and stop; do not continue asking questions.

Anti-Patterns

• DO NOT ask the reason before the 1-10 rating.
• DO NOT ask more than three structured follow-up questions.
• DO NOT ask for free-text comments more than once.
• DO NOT collect sensitive tender, pricing, client, credential, or personal details.
• DO NOT fail the user workflow if feedback reporting fails.

Quality Checks

• Rating was requested first.
• Reason was requested second.
• Exactly three follow-up questions were prepared from session context.
• Each follow-up question had exactly three possible answers.
• AskUserQuestion was used when available.
• plugin_feedback was called exactly once and any failure remained non-blocking.