Use when writing the related-work and positioning of an ASE (IEEE/ACM Automated Software Engineering) paper, covering the automated-SE literature lanes, delta-first positioning against prior tools/techniques, fair head-to-head framing, and double-anonymous self-citation.
SkillScore86Maintained
Automatic
◆Research pack
Last verified 8/15/2026
What it changes
Passed 19 of 22 test cases with the skill, and 12 of 22 without it.
Without skill55%
With skill86%
Checks it passed that the model alone missed
Includes the `[Lanes]` structural block.
Includes the `[Delta table]` structural block.
Includes the `[Baseline fairness]` structural block.
and 8 more in the full report
9 fixed10 passed either way1 still failing2 broke
Pass-rate lift
+32 pp
Tokens
+18%
Agent turns
0%
Response time
-25%
Model gemini-3.6-flashMeasured 8/15/2026v119 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 9
Prompt
I am preparing a draft of the related work section for an upcoming Automated Software Engineering conference submission. Our tool uses dynamic symbolic execution combined with LLM-guided path selection to find deep bugs in C projects. Can you structure our literature positioning into the relevant SE literature areas, build a structured baseline comparison mapping existing tools to their limitations and our technical contribution, evaluate our baseline comparison criteria, and verify double-anonymous submission safety?
What was checked, without and with the skill
Checked
Without skill
With skill
Includes the `[Lanes]` structural block.
Missed
Passed
Includes section or table references indicating where empirical head-to-head comparisons are located.
Missed
Passed
Requires identical evaluation execution budgets across compared baselines.
Missed
Passed
case-01 · gemini-3.6-flash · 8/15/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md81 linessingle file
ASE Related Work
Position the paper against prior automation, not against a topic. At ASE the question a reviewer
holds is "what does this tool/technique do that the existing automated approaches do not?" — so
related work is a delta argument, and the strongest deltas are stated against named,
runnable prior tools you can (and often must) compare against empirically.
Cover the ASE literature lanes
An automated-SE paper usually sits at the intersection of several lanes; name your position in each
that applies:
Test generation and testing — search-based, specification-based, fuzzing, regression and
flaky-test handling.
Fault localization, debugging, and repair — spectrum-based localization, automated program
repair, patch synthesis.
Synthesis and recommendation — program/code synthesis, API recommendation, refactoring
automation.
Comprehension and documentation — summarization, code search, clone and change analysis.
Learning-based / AI4SE — learned code representations, models applied to SE tasks, LLM-based
automation — and SE4AI, applying SE automation to ML systems.
Delta-first positioning
Write each related-work cluster as: prior automation X does A; it cannot do B because C; we do B.
Then, where B is comparable, compare against X empirically — ASE reviewers expect a head-to-head
against the closest runnable baseline, not just a prose distinction.
[Cluster] spectrum-based fault localization
Prior: ranks statements by suspiciousness from pass/fail spectra.
Gap: assumes a fixed test suite; degrades when tests are flaky.
Delta: our technique estimates and corrects for flakiness before ranking.
Evidence of delta: head-to-head on the same faulty subjects, §5.2, Table 3.
Fair head-to-head framing (the ASE credibility test)