nl-debugger

Files (1)

• SKILL.md

SKILL.md (first lines)

# SKILL: nl-debugger

## Purpose
Convert a user's natural-language issue into an executable debugging plan, including the most relevant downstream skills to run.

## Inputs
- user_message (string, required)

## Output
- inferred_issue_type
- extracted_entities
- missing_info
- immediate_checks (ordered, numbered list)
- next_skills (ordered, one line)
- expected_results

## Workflow
1. Identify the issue type: booking | member | access | wallet | resource | billing | schedule
2. Extract only explicitly stated entities (names, emails, dates, time ranges, resource names, locations)
3. List missing data required to run concrete queries (memberId, resourceId, date range, bookingId)
4. Produce 1-3 immediate checks as a numbered list
5. Recommend the downstream skills to execute (from your atomic skill set) in a single line

## Notes
- Keep outputs stable and structured for copy/paste.
- Do not infer dates or time ranges beyond what the user explicitly states.
- If dates are relative (e.g., "yesterday"), convert to an absolute date using the current locale date.
- Default time range: last 7 days through next 7 days only if the user asked for a range and did not provide one.
- Always return missing_info if any critical identifiers are absent.

From Raulitoflying/optix-graphql-mcp, MIT.

More from Raulitoflying/optix-graphql-mcp