Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create a new context file for a research Phase. Called once at Phase start to initialize the file that subsequent context-checkpoint calls will append to. Use this skill whenever a new research Phase begins and a fresh context file is needed.
.claude/skills/yogsoth-ai-context-init/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✓→✗ | ▼ Worse | -51% | 0% |
| case-12 | ✓→✗ | ▼ Worse | -19% | 0% |
| case-19 | ✓→✗ | ▼ Worse | -50% | 0% |
| case-20 | ✓→✓ | = Same ✓ | 135% | 0% |
| case-21 | ✓→✓ | = Same ✓ | 70% | 0% |
Create a new context file for a research Phase. One file per Phase — subsequent checkpoints append to this file.
Once at the beginning of each research Phase, imported by the plan file:
step: "import context-management:context-init"Get the current time:
bashdate +%Y-%m-%d-%H-%M
Output format: yyyy-mm-dd-hh-mm (e.g., 2026-05-16-14-30)
Based on the current Phase context, decide a topic slug:
lit-survey, gap-analysis, protein-folding-ideationCreate the file at: context/<timestamp>-<topic-slug>.md
Write the initial content:
markdown# <Phase Topic> > Created: <yyyy-mm-dd hh:mm> > Topic: <Research Topic> > Phase: <Phase Name> ## Plan Context <Plan excerpt for this Phase: objectives, expected outputs, constraints, relationship to other Phases in the pipeline>
The Plan Context section should contain the relevant plan excerpt describing this Phase's goals, expected deliverables, and how it connects to the broader research pipeline.
No mid-paragraph line breaks: Write each prose paragraph as a single continuous line. Do not insert newlines inside a paragraph to wrap it at a column width. Newlines are only for separating paragraphs, list items, headings, and fenced code blocks. This applies to the Plan Context written here and to all content appended later by context-checkpoint.
If context/INDEX.md does not exist, create it:
markdown# Context Index | File | Phase | Topic | Checkpoints | Last Updated | |------|-------|-------|-------------|--------------|
Add a new row for this file:
markdown| <filename> | <Phase Name> | <Research Topic> | 0 | <yyyy-mm-dd hh:mm> |
If a context file for the current Phase already exists (check INDEX.md for a matching Phase name created in the current session), skip creation and return the existing file path.
The file path of the created (or existing) context file. This path is used by subsequent context-checkpoint calls within the same Phase.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,240 | 6,193 | -33% | 1 | 1 | 0% | 1,536 | 883 | -43% | 0 | 0 | — |
case-02 | fail→fail | 7,238 | 6,729 | -7% | 1 | 1 | 0% | 1,281 | 1,146 | -11% | 0 | 0 | — |
case-03 | fail→fail | 11,114 | 6,271 | -44% | 1 | 1 | 0% | 1,988 | 1,010 | -49% | 0 | 0 | — |
case-04 | fail→fail | 9,361 | 9,539 | +2% | 1 | 1 | 0% | 1,671 | 1,341 | -20% | 0 | 0 | — |
case-05 | pass→fail | 12,250 | 4,718 | -61% | 1 | 1 | 0% | 1,680 | 829 | -51% | 0 | 0 | — |
case-06 | fail→fail | 6,152 | 6,750 | +10% | 1 | 1 | 0% | 1,020 | 969 | -5% | 0 | 0 | — |
case-07 | fail→fail | 7,381 | 6,028 | -18% | 1 | 1 | 0% | 1,199 | 893 | -26% | 0 | 0 | — |
case-08 | fail→fail | 2,799 | 8,271 | +195% | 1 | 1 | 0% | 347 | 1,774 | +411% | 0 | 0 | — |
case-09 | fail→fail | 4,528 | 6,098 | +35% | 1 | 1 | 0% | 759 | 833 | +10% | 0 | 0 | — |
case-10 | fail→fail | 9,228 | 5,673 | -39% | 1 | 1 | 0% | 1,433 | 925 | -35% | 0 | 0 | — |
case-11 | fail→fail | 9,517 | 6,203 | -35% | 1 | 1 | 0% | 1,672 | 937 | -44% | 0 | 0 | — |
case-12 | pass→fail | 7,044 | 6,342 | -10% | 1 | 1 | 0% | 1,172 | 946 | -19% | 0 | 0 | — |
case-13 | fail→fail | 10,376 | 5,274 | -49% | 1 | 1 | 0% | 1,835 | 888 | -52% | 0 | 0 | — |
case-14 | fail→fail | 11,026 | 6,177 | -44% | 1 | 1 | 0% | 1,878 | 1,025 | -45% | 0 | 0 | — |
case-15 | fail→fail | 6,675 | 5,125 | -23% | 1 | 1 | 0% | 1,062 | 923 | -13% | 0 | 0 | — |
case-16 | fail→fail | 8,809 | 7,151 | -19% | 1 | 1 | 0% | 1,400 | 972 | -31% | 0 | 0 | — |
case-17 | fail→fail | 9,150 | 6,802 | -26% | 1 | 1 | 0% | 1,528 | 1,146 | -25% | 0 | 0 | — |
case-18 | fail→fail | 4,908 | 4,503 | -8% | 1 | 1 | 0% | 797 | 785 | -2% | 0 | 0 | — |
case-19 | pass→fail | 11,982 | 6,827 | -43% | 1 | 1 | 0% | 2,138 | 1,063 | -50% | 0 | 0 | — |
case-20 | pass→pass | 6,171 | 11,079 | +80% | 1 | 1 | 0% | 1,012 | 2,379 | +135% | 0 | 0 | — |
case-21 | pass→pass | 11,284 | 15,009 | +33% | 1 | 1 | 0% | 1,953 | 3,323 | +70% | 0 | 0 | — |
case-22 | pass→pass | 7,955 | 11,277 | +42% | 1 | 1 | 0% | 1,333 | 2,445 | +83% | 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. 22 cases were attempted, and 5 counted toward the lift figure. The other 17 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. A headline lift is not published for this run.
Other measured skills in the registry, with their headline benchmark lift.