$npx -y skills add evanklem/evanflow --skill evanflow-qaConversational bug discovery → issue draft. Light listening, background exploration, scope assessment. Asks before gh issue create — never auto-files. Use when conducting a QA session, triaging user-reported issues, or filing bugs.
| 1 | # EvanFlow: QA |
| 2 | |
| 3 | |
| 4 | **Removed from upstream:** automatic `gh issue create`. Always ask first. |
| 5 | |
| 6 | ## Vocabulary |
| 7 | |
| 8 | See `evanflow` meta-skill. |
| 9 | |
| 10 | ## When to Use |
| 11 | |
| 12 | - User describes a bug in their own words |
| 13 | - QA session walking through a feature |
| 14 | - Triaging multiple reports |
| 15 | |
| 16 | ## The Flow |
| 17 | |
| 18 | ### 1. Listen Lightly |
| 19 | |
| 20 | Let the user describe in their own words. **At most 2-3 short clarifying questions**, focused on: |
| 21 | |
| 22 | - What did you expect to happen? |
| 23 | - What actually happened? |
| 24 | - What were the steps? |
| 25 | |
| 26 | Don't pre-debug. Don't pre-judge. |
| 27 | |
| 28 | ### 2. Explore in Background (optional) |
| 29 | |
| 30 | Spin up an Agent (Explore type) to read the relevant codebase area. Goal: understand domain language and feature intent — **not to find a fix**. Fixes come later via `evanflow-debug`. |
| 31 | |
| 32 | ### 3. Assess Scope |
| 33 | |
| 34 | Decide: is this one issue, or does it need breakdown into multiple independent concerns that different contributors could tackle simultaneously? |
| 35 | |
| 36 | If multiple: each becomes its own issue. Mark blocking relationships honestly (don't serialize when work could parallelize). |
| 37 | |
| 38 | ### 4. Draft the Issue(s) |
| 39 | |
| 40 | **Show the draft to the user. ASK before running `gh issue create`.** |
| 41 | |
| 42 | Single-issue template: |
| 43 | |
| 44 | ```markdown |
| 45 | ## What happened |
| 46 | <one-paragraph user-facing description, no file paths> |
| 47 | |
| 48 | ## What I expected |
| 49 | <one-paragraph> |
| 50 | |
| 51 | ## Steps to reproduce |
| 52 | 1. ... |
| 53 | 2. ... |
| 54 | |
| 55 | ## Additional context |
| 56 | <anything that helps an engineer reproduce or scope> |
| 57 | ``` |
| 58 | |
| 59 | Breakdown template (parent + children): |
| 60 | |
| 61 | ```markdown |
| 62 | ## Parent issue |
| 63 | References: #<parent-id> |
| 64 | |
| 65 | ## What's wrong |
| 66 | <scoped to this child> |
| 67 | |
| 68 | ## What I expected |
| 69 | ... |
| 70 | |
| 71 | ## Steps to reproduce |
| 72 | ... |
| 73 | |
| 74 | ## Blocked by |
| 75 | - #<id> (if any) |
| 76 | |
| 77 | ## Additional context |
| 78 | ... |
| 79 | ``` |
| 80 | |
| 81 | ### 5. File (with permission) |
| 82 | |
| 83 | After user confirms each draft: `gh issue create --title "..." --body "$(cat <<'EOF' ... EOF)"`. |
| 84 | |
| 85 | Do NOT batch-file. Confirm each one. |
| 86 | |
| 87 | ### 6. Iterate |
| 88 | |
| 89 | Process each issue independently. The user signals when the QA session is done. |
| 90 | |
| 91 | ## Hard Rules |
| 92 | |
| 93 | - **No file paths or line numbers in issue bodies.** They become stale. |
| 94 | - **Use the project's domain language** (`CONTEXT.md`). |
| 95 | - **Describe behaviors, not code mechanics.** |
| 96 | - **Reproduction steps mandatory.** |
| 97 | - **Concise — readable in 30 seconds.** |
| 98 | - **Many thin issues > few thick issues.** |
| 99 | - **Mark blocking honestly to maximize parallelism.** |
| 100 | - **ALWAYS ask before `gh issue create`.** Never auto-file. |
| 101 | - **Never auto-commit.** This skill doesn't modify code anyway, but if it ever does — ask first. |
| 102 | |
| 103 | ## Hand-offs |
| 104 | |
| 105 | - Issue filed, want to debug it now → `evanflow-debug` |
| 106 | - Issue reveals a feature gap, not a bug → `evanflow-brainstorming` (or `evanflow-prd` if substantial) |