Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Query previous pi sessions to retrieve context, decisions, code changes, or other information. Use when you need to look up what happened in a parent session or any other session file.
.claude/skills/dicklesworthstone-session-query/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | 4% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 25% | 0% |
| case-23 | ✓→✓ | ＝ Same ✓ | 48% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -29% | 0% |
Query pi session files to retrieve context from past conversations.
This skill is automatically invoked in handed-off sessions when you need to look up details from the parent session.
Use the session_query tool:
session_query(sessionPath, question)sessionPath: Full path to the session file (provided in the "Parent session:" line)question: Specific question about that session (e.g., "What files were modified?" or "What approach was chosen?")session_query("/path/to/session.jsonl", "What files were modified?")
session_query("/path/to/session.jsonl", "What approach was chosen for authentication?")
session_query("/path/to/session.jsonl", "Summarize the key decisions made")The tool loads the session and uses an LLM to answer your question based on its contents. Ask specific questions for best results.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,021 | 15,592 | +73% | 1 | 1 | 0% | 624 | 442 | -29% | 0 | 0 | — |
case-06 | fail→fail | 42,096 | 14,851 | -65% | 1 | 1 | 0% | 4,683 | 426 | -91% | 0 | 0 | — |
case-02 | fail→fail | 48,754 | 15,467 | -68% | 1 | 1 | 0% | 7,081 | 420 | -94% | 0 | 0 | — |
case-03 | fail→fail | 12,980 | 14,074 | +8% | 1 | 1 | 0% | 1,290 | 367 | -72% | 0 | 0 | — |
case-04 | fail→fail | 13,396 | 16,658 | +24% | 1 | 1 | 0% | 1,507 | 524 | -65% | 0 | 0 | — |
case-05 | fail→pass | 13,460 | 11,870 | -12% | 1 | 1 | 0% | 1,395 | 1,341 | -4% | 0 | 0 | — |
case-07 | fail→fail | 23,571 | 14,719 | -38% | 1 | 1 | 0% | 3,130 | 430 | -86% | 0 | 0 | — |
case-08 | fail→fail | 15,896 | 14,132 | -11% | 1 | 1 | 0% | 355 | 395 | +11% | 0 | 0 | — |
case-09 | fail→fail | 15,366 | 13,904 | -10% | 1 | 1 | 0% | 322 | 434 | +35% | 0 | 0 | — |
case-10 | fail→fail | 31,993 | 14,180 | -56% | 1 | 1 | 0% | 260 | 428 | +65% | 0 | 0 | — |
case-11 | fail→fail | 9,452 | 13,977 | +48% | 1 | 1 | 0% | 555 | 406 | -27% | 0 | 0 | — |
case-12 | fail→fail | 15,645 | 14,157 | -10% | 1 | 1 | 0% | 267 | 389 | +46% | 0 | 0 | — |
case-13 | fail→fail | 47,099 | 14,299 | -70% | 1 | 1 | 0% | 5,155 | 423 | -92% | 0 | 0 | — |
case-14 | fail→fail | 15,002 | 14,459 | -4% | 1 | 1 | 0% | 278 | 425 | +53% | 0 | 0 | — |
case-15 | fail→fail | 12,053 | 17,337 | +44% | 1 | 1 | 0% | 1,201 | 427 | -64% | 0 | 0 | — |
case-16 | fail→fail | 12,090 | 14,385 | +19% | 1 | 1 | 0% | 1,190 | 391 | -67% | 0 | 0 | — |
case-17 | fail→fail | 17,140 | 14,639 | -15% | 1 | 1 | 0% | 2,891 | 393 | -86% | 0 | 0 | — |
case-18 | fail→fail | 14,796 | 14,648 | -1% | 1 | 1 | 0% | 225 | 403 | +79% | 0 | 0 | — |
case-19 | fail→fail | 21,701 | 14,242 | -34% | 1 | 1 | 0% | 297 | 423 | +42% | 0 | 0 | — |
case-20 | pass→pass | 11,803 | 10,749 | -9% | 1 | 1 | 0% | 1,191 | 1,236 | +4% | 0 | 0 | — |
case-21 | fail→fail | 7,686 | 23,969 | +212% | 1 | 1 | 0% | 252 | 1,103 | +338% | 0 | 0 | — |
case-22 | pass→pass | 11,765 | 12,686 | +8% | 1 | 1 | 0% | 1,157 | 1,443 | +25% | 0 | 0 | — |
case-23 | pass→pass | 12,669 | 16,389 | +29% | 1 | 1 | 0% | 1,575 | 2,327 | +48% | 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. 23 cases were attempted, and 4 counted toward the lift figure. The other 19 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 +4 percentage points is the difference between those two pass rates over the 4 comparable cases. 1 case got worse with the skill loaded, and it is 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.