Use when packaging the code, datasets, knowledge graphs, prompts, and demo systems around a CIKM paper — choosing the artifact form per track (research, applied, resource, demo), meeting the resource track's reuse-and-documentation bar, and staging anonymous review artifacts into citable public releases.
SkillScore68Maintained
Automatic
◆Research pack
Last verified 8/31/2026
What it changes
Passed 14 of 21 test cases with the skill, and 4 of 21 without it.
Without skill18%
With skill68%
Checks it passed that the model alone missed
Uses the exact line prefix '[Staging]'
Uses the exact line prefix '[KG fields]'
Uses the exact line prefix '[Track role]'
and 9 more in the full report
11 fixed4 passed either way7 still failing0 broke
Pass-rate lift
+50 pp
Tokens
+11%
Agent turns
0%
Response time
-33%
Model gemini-3.6-flashMeasured 8/31/2026v115 of 22 cases passed with skill · 21 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 11 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 11
Prompt
Our CIKM dataset repo contains high-quality parquet files, comprehensive field datasheets, and Python baseline scripts, but we left out a LICENSE file because legal counsel is still reviewing terms. Evaluate repository readiness.
What was checked, without and with the skill
Checked
Without skill
With skill
Identifies an unlicensed dataset as unadoptable
Passed
Passed
Uses the exact line prefix '[Adoption blocker]'
Missed
Passed
case-18 · gemini-3.6-flash · 8/31/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md114 linessingle file
CIKM Artifact Evaluation
CIKM has no verified badge-granting artifact track (2026 lineup: full, short,
applied, resource, demo — source map, 2026-07-08), but artifacts do more work here
than at most venues, because two of the five tracks are about artifacts: Resource
papers publish them, and Demonstration papers perform them. This skill covers the
artifact strategy per track and the staging from anonymous review object to public
citable release.
Artifact role by track
Track
The artifact is...
Bar it must clear
Full / Short Research
Supporting evidence, cited anonymously
Regenerates headline results; survives a ten-minute inspection
Applied Research
Proof of practice
Substantiates the launch/deployment claim the track requires
Resource
The contribution itself
Documented, licensed, maintained, adoptable by strangers
Demonstration
The thing being demonstrated
Runs live, in front of people, on conference-venue conditions
Resource-track packaging bar
A CIKM resource paper is graded on whether others will build on the object. The
verified call names data resources, software, evaluation tasks, and open-source
frameworks. Package accordingly:
Provenance and license first: where the data came from, what consent/terms
cover it, and a license a university lab and a company can both act on. An
unlicensed dataset is unadoptable regardless of quality.
Schema and statistics: field-level documentation, size, class balance, known
gaps — the datasheet discipline, applied without the buzzword.
Baselines included: a resource ships with reference numbers and the scripts
that produce them, or every adopter re-invents the evaluation differently.
Maintenance posture: versioning scheme, issue channel, and an honest
statement of how long the hosting is funded. Reviewers of resource papers have
watched too many datasets rot at dead URLs.
All of it inside 4 pages (2026) — the paper is the resource's brochure and
datasheet condensed; depth lives in the repository.
KG-specific packaging
Knowledge graphs — a CIKM staple — need artifact fields other objects skip: the