Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Gather phase context through adaptive questioning before planning. Use --all to skip area selection and discuss all gray areas interactively. Use --auto to skip interactive questions (Claude picks recommended defaults). Use --chain for interactive discuss followed by automatic plan+execute. Use --power for bulk question generation into a file-based UI (answer at your own pace).
.claude/skills/davepoon-gsd-discuss-phase/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-13 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-20 | ✗→✓ | ▲ Improved | -25% | 0% |
<objective> Extract implementation decisions that downstream agents need — researcher and planner will use CONTEXT.md to know what to investigate and what choices are locked.
How it works:
Output: {phase_num}-CONTEXT.md — decisions clear enough that downstream agents can act without asking the user again </objective>
<execution_context> @${CLAUDE_PLUGIN_ROOT}/workflows/discuss-phase.md @${CLAUDE_PLUGIN_ROOT}/workflows/discuss-phase-assumptions.md @${CLAUDE_PLUGIN_ROOT}/workflows/discuss-phase-power.md @${CLAUDE_PLUGIN_ROOT}/templates/context.md </execution_context>
<runtime_note> Copilot (VS Code): Use vscode_askquestions wherever this workflow calls AskUserQuestion. They are equivalent — vscode_askquestions is the VS Code Copilot implementation of the same interactive question API. </runtime_note>
<context> Phase number: $ARGUMENTS (required)
Context files are resolved in-workflow using init phase-op and roadmap/state tool calls. </context>
<process> Mode routing:
bashDISCUSS_MODE=$(gsd-sdk query config-get workflow.discuss_mode 2>/dev/null || echo "discuss")
If DISCUSS_MODE is "assumptions": Read and execute @${CLAUDE_PLUGIN_ROOT}/workflows/discuss-phase-assumptions.md end-to-end.
If DISCUSS_MODE is "discuss" (or unset, or any other value): Read and execute @${CLAUDE_PLUGIN_ROOT}/workflows/discuss-phase.md end-to-end.
MANDATORY: The execution_context files listed above ARE the instructions. Read the workflow file BEFORE taking any action. The objective and success_criteria sections in this command file are summaries — the workflow file contains the complete step-by-step process with all required behaviors, config checks, and interaction patterns. Do not improvise from the summary. </process>
<success_criteria>
</success_criteria>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 4,248 | 3,705 | -13% | 1 | 1 | 0% | 184 | 843 | +358% | 0 | 0 | — |
case-03 | fail→fail | 15,890 | 4,560 | -71% | 1 | 1 | 0% | 2,827 | 874 | -69% | 0 | 0 | — |
case-01 | fail→fail | 13,607 | 4,866 | -64% | 1 | 1 | 0% | 2,360 | 897 | -62% | 0 | 0 | — |
case-04 | pass→fail | 26,141 | 6,366 | -76% | 1 | 1 | 0% | 4,883 | 939 | -81% | 0 | 0 | — |
case-05 | fail→fail | 3,587 | 8,853 | +147% | 1 | 1 | 0% | 590 | 2,184 | +270% | 0 | 0 | — |
case-06 | fail→fail | 10,127 | 4,792 | -53% | 1 | 1 | 0% | 1,381 | 901 | -35% | 0 | 0 | — |
case-07 | pass→fail | 10,118 | 7,838 | -23% | 1 | 1 | 0% | 1,725 | 916 | -47% | 0 | 0 | — |
case-08 | fail→fail | 5,169 | 5,087 | -2% | 1 | 1 | 0% | 855 | 938 | +10% | 0 | 0 | — |
case-09 | fail→fail | 4,981 | 5,214 | +5% | 1 | 1 | 0% | 813 | 979 | +20% | 0 | 0 | — |
case-10 | fail→fail | 10,585 | 4,630 | -56% | 1 | 1 | 0% | 1,710 | 866 | -49% | 0 | 0 | — |
case-11 | pass→pass | 10,035 | 6,147 | -39% | 1 | 1 | 0% | 1,584 | 1,634 | +3% | 0 | 0 | — |
case-12 | pass→pass | 10,884 | 4,796 | -56% | 1 | 1 | 0% | 1,744 | 1,507 | -14% | 0 | 0 | — |
case-13 | fail→pass | 6,809 | 1,292 | -81% | 1 | 1 | 0% | 1,211 | 806 | -33% | 0 | 0 | — |
case-14 | fail→pass | 7,743 | 2,079 | -73% | 1 | 1 | 0% | 1,301 | 1,024 | -21% | 0 | 0 | — |
case-15 | pass→pass | 7,278 | 3,255 | -55% | 1 | 1 | 0% | 1,283 | 1,153 | -10% | 0 | 0 | — |
case-16 | fail→pass | 11,203 | 9,087 | -19% | 1 | 1 | 0% | 1,834 | 2,064 | +13% | 0 | 0 | — |
case-17 | fail→fail | 10,459 | 1,554 | -85% | 1 | 1 | 0% | 1,929 | 902 | -53% | 0 | 0 | — |
case-18 | pass→fail | 6,496 | 2,177 | -66% | 1 | 1 | 0% | 1,066 | 1,018 | -5% | 0 | 0 | — |
case-19 | fail→pass | 11,282 | 3,777 | -67% | 1 | 1 | 0% | 1,875 | 1,291 | -31% | 0 | 0 | — |
case-20 | fail→pass | 6,747 | 1,346 | -80% | 1 | 1 | 0% | 1,131 | 846 | -25% | 0 | 0 | — |
case-21 | pass→pass | 9,073 | 5,155 | -43% | 1 | 1 | 0% | 1,520 | 1,503 | -1% | 0 | 0 | — |
case-22 | fail→pass | 4,788 | 2,157 | -55% | 1 | 1 | 0% | 814 | 983 | +21% | 0 | 0 | — |
case-23 | pass→pass | 9,594 | 6,279 | -35% | 1 | 1 | 0% | 1,541 | 1,706 | +11% | 0 | 0 | — |
case-24 | pass→pass | 8,777 | 6,977 | -21% | 1 | 1 | 0% | 1,562 | 1,377 | -12% | 0 | 0 | — |
case-25 | fail→pass | 8,085 | 2,943 | -64% | 1 | 1 | 0% | 1,306 | 1,174 | -10% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted, and 16 counted toward the lift figure. The other 9 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +16 percentage points is the difference between those two pass rates over the 16 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.