mcpField
C
context-budget-guard

context-budget-guard

github.com/archieindianAgent workflows7351 installsUpdated 26w ago
Repository

About context-budget-guard

context-budget-guard is an OpenClaw skill for tracking context use during long sessions and multi-stage work. It prompts context compaction when estimated usage exceeds 70%.

Key features

  1. 01Initialize and maintain session state
  2. 02Estimate context use at task checkpoints
  3. 03Trigger context management above 70%
  4. 04Record compaction details and session end

Use cases

  1. 01Monitor context during a long session
  2. 02Check context before a major workflow step
  3. 03Compact before loading large files or subagents

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/archieindian/openclaw-superpowers --skill context-budget-guard

After installing, type/context-budget-guardin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

check.py
example-state.yaml
STATE_SCHEMA.yaml

Details

From the project README

Context Budget Guard

State file: ~/.openclaw/skill-state/context-budget-guard/state.yaml

Don't wait for the model to go incoherent. Act at 70%, not 95%.

When to Use

  • At the start of any long-running session (>30 min expected)
  • Before each major task step in a multi-hour workflow
  • When long-running-task-management advances to a new stage

The Process

Step 1: Initialize (Session Start)

Write to state: session_start timestamp, compaction_count: 0, status: monitoring, threshold_pct: 70.

Step 2: Check Budget (Before Each Major Step)

Estimate current context usage as a rough percentage:

  • Low (<50%) — continue normally
  • Medium (50–70%) — note in state, proceed with caution (avoid loading large files)
  • High (>70%) — trigger context-window-management NOW, before continuing

To estimate: count approximate tokens from recent messages, loaded file contents, and active task context.

Step 3: After Compaction
  • Increment compaction_count in state
  • Write last_compacted_at, strategy_used (passed from context-window-management)
  • Reset mental estimate to ~20% (post-compaction baseline)
Step 4: Session End

Update state: status: idle, session_end timestamp.

Key Principles

  • The 70% threshold is the trigger — earlier is always better than later
  • Check budget BEFORE loading large files or running subagents, not after
  • If in doubt: compact. A clean context costs less than a confused one.
  • Works best paired with long-running-task-management — check budget at every checkpoint

Comments

Markdown supported