Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
crispy13 avatar

Repo Memory Distill

  • 1 installs
  • Updated April 7, 2026
  • crispy13/vardict_rs

Distill repository memory for variant dictionary analysis using Rust performance optimization.

About

VarDict Rust is a Rust port of VarDictJava for variant calling with improved performance. Use for genomic analysis with bioinformatics tools.

  • Rust port of VarDictJava
  • Performance parity benchmarking

Repo Memory Distill by the numbers

  • 1 all-time installs (skills.sh)
  • Ranked #105 of 121 Rust skills by installs in the Skillselion catalog
  • Data as of Jul 8, 2026 (Skillselion catalog sync)
npx skills add https://github.com/crispy13/vardict_rs --skill repo-memory-distill

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs1
Last updatedApril 7, 2026
Repositorycrispy13/vardict_rs

What it does

Distill repository memory for variant dictionary analysis using Rust performance optimization.

Files

SKILL.mdMarkdownGitHub ↗

Repo Memory Distill

Purpose

Consolidate accumulated /memories/repo/ files into thematic summary files for better scanability.

When to Trigger

Trigger this workflow when /memories/repo/ exceeds about 50 files. The current project has 59 files, so this skill is intended for the current scale of repository memory.

Procedure

Step 1: Inventory

  • List all files in /memories/repo/.
  • Read each file to understand its content.
  • Note the date from the filename suffix, for example _20260310, when present.

Step 2: Categorize by Theme

Adapt the categories to the actual content, but start from this mapping:

CategoryFileContent Types
Parity Trapsparity-traps.mdBugs that fooled diagnosis, misleading symptoms, false leads
Debugging Patternsdebugging-patterns.mdEffective debugging techniques, diagnostic approaches
Design Decisionsdesign-decisions.mdArchitecture choices, why X over Y, API decisions
Known Limitationsknown-limitations.mdAccepted divergences, platform-specific issues
Infrastructureinfrastructure.mdBuild system, conda, test harness, CI notes
SV and Realignersv-and-realigner.mdStructural variant processor and realigner specifics
VecMap and Collectionsvecmap-and-collections.mdCollection type decisions, ordering issues
Memory and Performancememory-and-performance.mdMemory optimization, allocation, benchmarking

Keep any files that do not fit these themes in their original form until a better category is clear.

Step 3: Draft Summaries

For each category:

  • Create a new summary file with a clear header.
  • Group entries by sub-topic.
  • Keep each entry to 1-3 bullet points max.
  • Preserve the original date.
  • Preserve unique insights and never discard information.
  • Use this entry format:
### <Topic> (YYYY-MM-DD)
- Key finding 1
- Key finding 2
- Related: <other entries or files if cross-referenced>

If a source note has no date suffix, keep the topic but mark the date conservatively, for example using the known file date if available elsewhere or undated if it cannot be recovered safely.

Step 4: User Review

  • Present the draft summaries to the user before deleting any originals.
  • Show a mapping of which original files were merged into which summary.
  • Get explicit approval before proceeding to deletion.

Step 5: Replace

  • Create the new summary files in /memories/repo/.
  • Delete only the original files that were merged and approved for removal.
  • Keep any files that do not fit into categories as-is.

Quality Rules

  • Never discard unique insights. If in doubt, keep the entry.
  • Preserve original dates in each entry.
  • Keep entries concise: 1-3 bullets per topic.
  • Cross-reference related entries within summaries.
  • If two entries contradict each other, keep both and note the contradiction.
  • Reduce file count substantially, but keep the summaries lossless in information.

Constraints

  • Always get user approval before deleting any memory files.
  • Never merge files across memory scopes: repo, session, and user memories must stay separate.
  • Keep the total summary file count between 5 and 12, depending on content diversity.
  • Each summary file should be scannable in under 30 seconds.

Related skills

Rustbackend

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.