From tl
Use when you've found a defect and need to file it. Creates a structured bug report with repro steps and priority from a description or prior /critique or /review output. Keywords: bug, report, issue, defect, file, submit, track, broken.
How this skill is triggered — by the user, by Claude, or both
Slash command
/tl:bug <description of the bug><description of the bug>This skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are running the **bug** skill — filing a structured bug report in your task tracker. Input: **$ARGUMENTS**
You are running the bug skill — filing a structured bug report in your task tracker. Input: $ARGUMENTS
Detect input [pipe-format or direct description]
-> Extract details [artifact, symptom, repro, severity]
-> Confirm with user [priority, scope]
-> File in task tracker
-> Report [pipe format]
Detect upstream pipe-format output in context. Prioritize items tagged as FLAW, RISK, GAP, or with severity HIGH/CRITICAL.
State what you found: "Reading N items from /skill-name output above" or "Working from direct description."
For each bug candidate, determine:
| Field | How to Extract |
|---|---|
| Artifact | Which skill, agent, rule, hook, or file is affected? Search the codebase if unclear. |
| Symptom | What goes wrong? Incorrect output, crash, silent failure, missing behavior? |
| Repro steps | What sequence triggers the bug? Include the skill invocation or input if known. |
| Expected behavior | What should happen instead? |
| Severity | P0 (crash/data loss), P1 (wrong output), P2 (minor/cosmetic), P3 (edge case) |
If extracting from pipe-format, map upstream metadata:
If details are ambiguous, ask the user rather than guessing.
Present the extracted bug report to the user for confirmation before creating the issue:
Bug: <title>
Artifact: <file or skill affected>
Severity: P<N>
Symptom: <what goes wrong>
Repro: <steps to trigger>
Expected: <correct behavior>
If multiple bugs were extracted from pipe-format input, present all of them and ask the user which to file. Do not silently file all of them.
For each confirmed bug, create a tracked issue with:
<affected artifact>: <concise symptom> (e.g., "/gather: silently drops items past 20 results")If the bug relates to an existing epic or parent task, link it as a child.
Emit the result in pipe format:
## Filed Bug Reports
**Source**: /bug
**Input**: <one-line summary of input>
**Pipeline**: <upstream chain -> /bug (N items)> or (none — working from direct input)
### Items (N)
1. **<task-id>: <title>** — P<N>, <artifact>
- source: <file path or skill name>
### Summary
<One paragraph: what was filed, severity distribution, recommended next action (e.g., "run /sprint to dispatch fixes" or "add to next consolidation pass").>
npx claudepluginhub tyevans/tackline --plugin tacklineTurns raw bug reports into groomed, fixable tasks with reproducers, expected-vs-actual, and regression-test acceptance criteria. Use when a bug report lacks detail or a flaky test needs a proper task.
Files a GitHub bug issue from a description: structures a clear report (summary, steps, expected vs actual, environment), previews it, then creates the issue with the Bug issue type or bug label.
Captures user-reported bugs as durable GitHub issues using the project's domain language. Avoids file paths/line numbers; checks for duplicates.