token-budgeted-compaction-with-reinjection-a34bde — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited token-budgeted-compaction-with-reinjection-a34bde (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Domain: compaction Trigger: Use when session history is growing toward the API limit and you need a safe way to trim older context while still reinjecting critical artifacts like skill metadata. Source Pattern: Distilled from reviewed session memory, compaction, and context-budgeting implementations.
Run compaction in capped batches tied to a per-skill token budget: truncate each skill/document to <= 5k tokens, enforce an overall skill budget (25k tokens), bubble up warnings when budgets are hit, and only reinject the trimmed content when downstream turns request it. Keep a reusable helper that resets the per-skill counters after a compact run so the next turn starts from a clean slate.
When a Claude Code conversation runs long, apply this skill to compact turns by triming per-skill summaries down to 5k tokens, track the remaining budget, warn when the budget is exhausted, and include the reinjected skill attachments alongside the next query so the model has the context it needs.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.