loading
loading
Keep long-running agents under token limits without losing critical state.
The fix for the wall every long-running agent hits — it summarizes old turns before the context window blows, instead of crashing mid-task.
Listed for review
No verified public repo for this skill yet, so this page does not give you an install command. Skills with a verified source install in one command — or fully manual: copy the skill folder into .claude/skills/ and your agent picks it up.
Boostor Quality Score
84/100 · B
Context Budgeter instruments your agent loop to measure token usage per turn and automatically summarizes older conversation segments when approaching the context window ceiling. It exposes a priority queue API so you can mark certain messages as eviction-safe. Works with Claude Code, Codex, and any LangGraph workflow.
Review an agent's logging so a production failure can be replayed without exposing secrets.
Sets the bar at replaying a real failure from the trace without leaking secrets — and catches teams logging cost but not per-step cause.
Transparent + deterministic: every point above is computed from this skill's real fields plus a prompt-injection safety scan. No black box, no pay-to-rank.
Scaffold, test, and publish a Model Context Protocol server in minutes.
Writing an MCP server by hand means hours of boilerplate before the first tool runs. This validates against the spec so you skip the silent-failure stage.
Quality-gate an MCP server for schema clarity, tool risk, auth, error behavior, and smoke tests.
More than a schema check: it gates every MCP tool on inputs, failure behavior, and auth — and catches injection riding in through tool descriptions.