Skip to content
OpenAgentSkill
Browsing is open · Catalog captured 2026-11-21 · Automatic updates and account services are paused.
Development

bug-echo

After fixing a bug, find and rate other instances of the same pattern in the codebase. Two modes: described, or inferred from a recent fix with self-validation. Triggers: "run bug-echo", "echo this fix", "scan for similar bugs", "find other instances", "after-fix scan".

Terry Nyberg, Coffee & Code LLC · Static checks on file

Installation

Read the source and dependencies before running a command. Copying does not install or execute anything.

npx skills add Terryc21/bug-echo --skill bug-echo

About this skill


name: bug-echo description: 'After fixing a bug, find and rate other instances of the same pattern in the codebase. Two modes: described, or inferred from a recent fix with self-validation. Triggers: "run bug-echo", "echo this fix", "scan for similar bugs", "find other instances", "after-fix scan".' license: Apache-2.0 allowed-tools: [Grep, Glob, Read, Write, Edit, Bash, AskUserQuestion, Agent] metadata: version: 1.7.0 author: Terry Nyberg, Coffee & Code LLC tier: execution category: debugging

bug-echo

Quick Ref: After a bug fix, identify the pattern, scan the codebase, classify findings, and produce a rated report. Output: .agents/research/YYYY-MM-DD-bug-echo-<slug>.md.

Best invoked after a real fix

bug-echo is most effective when the pattern came from a fix that just shipped. A real fix proves which anti-pattern matters in your specific codebase. Pattern matching after a real fix is dramatically more accurate than pattern matching from a theoretical catalog — the fix is the evidence the pattern is a bug.

The high-leverage loop is surface → verify → generalize, three skills working in sequence:

  1. Surface — /unforget (or any tracker) shows you a deferred row you're about to mark Fixed.
  2. Verify — Before trusting the closure, confirm the fix is real. /radar-suite focus on <symbol> (or just reading the file) catches stale Open rows where the fix shipped weeks ago and nobody updated the ledger.
  3. Generalize — Run /bug-echo with a one-sentence description of what the fix replaced. The output is a rated list of every echo of that anti-pattern across the codebase — including instances that haven't crashed yet but sit under the same runtime conditions.

Bugs that haven't fired yet are the highest-ROI thing in any audit cycle. They cost the same to fix as crashed bugs, but you skip the cost of the crash itself (lost user trust, support tickets, root-cause investigation under deadline). bug-echo is the systematic way to find them.

Companion skills:

  • unforget (https://github.com/Terryc21/unforget) — the surface; consolidates deferred work in one file
  • radar-suite (https://github.com/Terryc21/radar-suite) — the verifier; confirms the fix is real before bug-echo generalizes

bug-echo also runs standalone when you describe the pattern manually (Step 2A below) — useful when no recent fix exists but you've spotted a shape worth chasing.


This is a workflow to run, not a procedure to summarize. The user invoked it because they want the scan performed and the findings in hand; a description of what bug-echo would do leaves them exactly where they started.

Rate every BUG finding on all six dimensions — Urgency, Risk of Fixing, Risk of Not Fixing, ROI, Blast Radius, and Fix Effort — using the Issue Rating Table in Step 5. The six exist because "here are 14 bugs" is not actionable: the user has to decide what to fix before the next release and what to leave. A finding missing Risk of Not Fixing or Blast Radius can't be triaged against the others, so it silently drops out of that decision no matter how real the bug is.

Narrate the user experience for every BUG and WATCH finding — what a person hits while the bug is live, and what they get instead once it is fixed. This is required in all three report modes (inline, full, and any promoted finding), in plain user-facing language: what they see, tap, and feel, not the mechanism. The six ratings say how bad a finding is on technical axes; none of them says what it costs the person using the app, which is the axis the fix-or-defer call is actually made on.

Sibling findings make this especially easy to get wrong, because their user cost is often not the same as the original fix's. The same anti-pattern in a rarely-reached settings pane and in the primary add-item flow are one pattern and two very different experiences. Narrate each site's own cost rather than restating the original's.

Two rules keep it honest:

  • 🛑 Findings with no visible symptom still get the field. Say so explicitly and name what the user is protected from — a crash that never happens, data that is not silently lost. Invisible wins are the ones most often dropped and frequently the most valuable. "No visible change; prevents X" is a complete answer.
  • 🛑 Never invent a symptom to fill the field. If you cannot say what the user would experience, that is evidence the finding may be theoretical — reclassify it to REVIEW or say plainly that the impact is unclear. An invented symptom inflates apparent severity and corrupts the triage the ratings exist to support.

This skill uses Claude's native tools only. No external scripts or pattern catalogs. AST-grep is optional; if it is installed, prefer it for higher precision on Swift, otherwise fall back to regex via the Grep tool.


Pre-flight

Before any scanning work, verify the working environment is sane.

  1. Check for uncommitted changes:

    • Run git status --porcelain via Bash.
    • If output is non-empty (uncommitted changes exist), use AskUserQuestion to ask: "There are uncommitted changes. If these are from a prior bug-echo session, commit them first so this run has a clean baseline. Otherwise: commit before scanning, or proceed anyway?" Options: "Commit first", "Proceed (accept risk)", "Cancel". On "Commit first", show files changed and stop with a request that the user commit. On "Cancel", stop. On "Proceed (accept risk)", log "User accepted risk of uncommitted changes" and continue.
  2. Note build manifest presence (advisory):

    • Detect via Glob whether Package.swift, xcodeproj, Cargo.toml, package.json, or a similar build manifest exists in the project root. Note the result in the report header. Do NOT run a build — that's the user's responsibility before applying any fix this skill suggests. If no manifest is detected, mention it but continue scanning. This step is advisory metadata for the r
Read the complete source ↗