Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when positioning a CSCW paper's related work — engaging the venue's own concept lineage, the PACMHCI-era literature, neighboring HCI venues, and the social sciences CSCW borrows from, with era-correct citations and anonymity-safe self-references for journal-model review.
.claude/skills/brycewang-stanford-cscw-related-work/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 54% | 0% |
Related work at CSCW is not a defensive perimeter; it is where the paper joins a forty-year conversation. Reviewers here are unusually likely to be the authors of the traditions you cite, to know the concept lineage personally, and — because of Revise and Resubmit — to check in round two whether the engagement you added in round one is real.
| Literature | What reviewers check | Failure smell | | --- | --- | --- | | CSCW's own lineage | That you know the concept history (awareness, articulation work, incentive misalignment, place/space) before renaming a piece of it | "We introduce the novel notion of..." for an idea with a 1992 citation | | Recent PACMHCI-era CSCW work | Engagement with the last ~3-5 years of the venue on your topic | A related-work section whose newest CSCW citation predates the journal era | | Neighboring venues | CHI, ICWSM, GROUP, DIS, and TOCHI work on the same phenomenon, cited by contribution rather than lumped | Treating a CHI interface study and a CSCW practice study as interchangeable | | Source disciplines | The sociology, organization studies, communication, or STS work your framing borrows | Invoking a social theory via a secondary HCI citation only |
Braiding matters more than coverage: a paragraph per literature, each ending in what this paper takes from or contests in it, beats an alphabetical tour.
CSCW's publication record has a seam in it, and sloppy citations tear on the seam:
No. CSCW, Article 31." There is no "Proceedings of CSCW 2019" for papers; a bibliography entry claiming one signals the authors have not read carefully.
conf/cscw stream) or the ACM DL PACMHCICSCW track; fix venue strings your reference manager guessed.
The Lasting Impact lineage in resources/exemplars/library.md doubles as a spine for concept-history paragraphs: if your topic touches failure-of-adoption, cite through Grudin 1988; awareness, through Dourish & Bellotti 1992; place and presence, through Harrison & Dourish 1996; sensitive-community storytelling, through Dimond et al. 2013; moderation at platform scale, through the PACMHCI-era trace canon.
name the condition that explains the difference; otherwise write "extending X to a setting where..."
interaction-level boundary of your claim; an ICWSM citation should mark the measurement baseline you build past.
analytic work in your paper, cite and quote the primary text; reviewers from the source discipline will notice a chain of second-hand paraphrases.
collaboration), state when the literature search closed — the R&R gap can span most of a year, and round-two reviewers may expect newer work acknowledged.
Submissions are anonymized. Your own prior work must be cited in the third person ("Building on the deployment reported by 12]...") and must not be the only scaffolding the argument stands on — a paper that cites one anonymous-adjacent thread six times deanonymizes itself structurally even with names removed. If prior work is on arXiv or a personal site, do not link it in a way that resolves identity.
text[Lineage] Oldest CSCW concept engaged: <paper, year> — engaged or name-dropped? [Recency] Newest CSCW/PACMHCI citation: <year>; search-close date stated? y/n [Neighbors] CHI/ICWSM/GROUP/DIS citations each given a boundary-marking job? y/n [Sources] Primary-text citation for each imported theory? y/n [Era check] All post-2017 CSCW refs cite PACMHCI vol/issue/article? y/n [Anonymity] Self-citations third-person, load-bearing count ≤ 2? y/n
Re-run the audit at every R&R: reviewers who asked for missing literature will read the new paragraphs first, and they can tell insertion from integration.
Citation-form facts verified against the ACM DL and dblp on 2026-07-08.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,945 | 26,153 | +54% | 1 | 1 | 0% | 2,766 | 4,027 | +46% | 0 | 0 | — |
case-02 | fail→pass | 26,429 | 24,262 | -8% | 1 | 1 | 0% | 4,205 | 3,935 | -6% | 0 | 0 | — |
case-03 | fail→pass | 17,993 | 13,306 | -26% | 1 | 1 | 0% | 2,139 | 3,027 | +42% | 0 | 0 | — |
case-04 | pass→pass | 15,134 | 14,216 | -6% | 1 | 1 | 0% | 2,871 | 2,576 | -10% | 0 | 0 | — |
case-05 | pass→pass | 17,618 | 12,923 | -27% | 1 | 1 | 0% | 2,666 | 2,791 | +5% | 0 | 0 | — |
case-06 | pass→pass | 17,079 | 17,616 | +3% | 1 | 1 | 0% | 2,103 | 2,817 | +34% | 0 | 0 | — |
case-07 | pass→pass | 11,736 | 14,158 | +21% | 1 | 1 | 0% | 1,933 | 2,656 | +37% | 0 | 0 | — |
case-08 | fail→pass | 19,215 | 17,887 | -7% | 1 | 1 | 0% | 2,337 | 3,202 | +37% | 0 | 0 | — |
case-09 | pass→pass | 18,459 | 18,411 | -0% | 1 | 1 | 0% | 2,162 | 3,202 | +48% | 0 | 0 | — |
case-10 | pass→fail | 12,666 | 13,155 | +4% | 1 | 1 | 0% | 2,050 | 3,242 | +58% | 0 | 0 | — |
case-11 | fail→fail | 16,403 | 23,531 | +43% | 1 | 1 | 0% | 2,496 | 4,255 | +70% | 0 | 0 | — |
case-12 | fail→pass | 24,769 | 25,239 | +2% | 1 | 1 | 0% | 2,735 | 4,200 | +54% | 0 | 0 | — |
case-13 | pass→pass | 18,034 | 16,552 | -8% | 1 | 1 | 0% | 2,139 | 3,929 | +84% | 0 | 0 | — |
case-23 | pass→pass | 15,888 | 9,275 | -42% | 1 | 1 | 0% | 2,486 | 1,592 | -36% | 0 | 0 | — |
case-14 | pass→pass | 18,186 | 20,150 | +11% | 1 | 1 | 0% | 2,198 | 3,758 | +71% | 0 | 0 | — |
case-15 | pass→pass | 11,548 | 8,374 | -27% | 1 | 1 | 0% | 1,087 | 2,609 | +140% | 0 | 0 | — |
case-16 | pass→pass | 22,394 | 16,598 | -26% | 1 | 1 | 0% | 2,916 | 3,597 | +23% | 0 | 0 | — |
case-17 | fail→pass | 19,609 | 20,247 | +3% | 1 | 1 | 0% | 2,216 | 3,147 | +42% | 0 | 0 | — |
case-18 | fail→pass | 37,215 | 7,771 | -79% | 1 | 1 | 0% | 2,819 | 2,294 | -19% | 0 | 0 | — |
case-19 | pass→pass | 14,754 | 8,037 | -46% | 1 | 1 | 0% | 1,537 | 1,731 | +13% | 0 | 0 | — |
case-20 | pass→pass | 20,901 | 23,564 | +13% | 1 | 1 | 0% | 2,683 | 3,971 | +48% | 0 | 0 | — |
case-21 | pass→pass | 18,010 | 14,501 | -19% | 1 | 1 | 0% | 2,284 | 3,726 | +63% | 0 | 0 | — |
case-22 | pass→pass | 25,039 | 23,224 | -7% | 1 | 1 | 0% | 2,857 | 4,206 | +47% | 0 | 0 | — |
case-24 | pass→pass | 15,735 | 14,494 | -8% | 1 | 1 | 0% | 1,865 | 2,727 | +46% | 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. 24 cases were attempted. The headline lift of +25 percentage points is the difference between those two pass rates over the 24 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.