Agent-Skills.md

Agent Skills: Context Management

Universal context management with proactive recommendations. Agent checks context levels and recommends context resets to users. Always active as foundation for context optimization.

UncategorizedID: tianjianjiang/smith/smith-ctx

Repository

tianjianjiang/smith
tianjianjiang

Install this agent skill to your local

pnpm dlx add-skill https://github.com/tianjianjiang/smith/tree/HEAD/smith-ctx

Skill Files

Browse the full folder contents for smith-ctx.

Download Skill

Loading file tree…

smith-ctx/SKILL.md

Skill Metadata

Name
smith-ctx
Description
Universal context-management foundation — proactive context-level checks and reset recommendations. Use when context usage grows high, when deciding whether to /clear or /compact, or when choosing what to retain across a context reset.

Context Management

Load if: Always active (context management foundation) Prerequisites: @smith-guidance/SKILL.md

Proactive Context Management

Agent role: Check context levels proactively, RECOMMEND actions to user.

Context lifecycle: explore → monitor → prepare (at warning threshold) → reset (at critical threshold)

To check context: Prompt "What is the current context usage?" to get percentage.

Agent RECOMMENDS - user executes the platform's context reset command.

Platform Reference

  • Claude Code: Warning 50%, Critical 60%, Targeted: "Summarize from here" (Esc+Esc), Action: /clear (stop hook enforced)
  • Kiro: Warning 70%, Critical 80%, Compact Auto, Clear New session
  • Cursor: Warning 70%, Critical 80%, Compact /summarize, Clear New chat

Progressive Disclosure

Loading order (cheapest first):

  1. Metadata scan: Glob/Grep for file locations
  2. Targeted read: Specific file sections only
  3. Full file: Only when actively modifying
  4. Broad explore: Delegate to subagent (isolated context)

Rules:

  • Filter with Grep before reading entire directories
  • Load targeted sections instead of full files when they suffice
  • Reuse context from files already read instead of re-reading them

Serena MCP Preference

When Serena MCP is available, prefer Serena tools over native tools:

Why: Kiro's readFile truncates, strReplace fails on duplicates. Serena's regex mode handles complex replacements reliably.

Tool preference:

  • Reading: search_for_pattern > find_symbol > native readFile
  • Writing: replace_content (regex) > replace_symbol_body > native strReplace
  • Context savings: 99%+ reduction with symbol-level operations

Information Retention

Always preserve:

  • Task goals
  • File:line locations
  • Architectural decisions
  • Incomplete work

Always discard:

  • Verbose tool outputs
  • Failed explorations
  • Redundant file reads

Reference format: Use file:line (e.g., auth.ts:234) instead of embedding content

Ralph Loop Context Management

Ralph burns ~1-3.5k tokens/iteration. At critical threshold, persist state before context reset: Serena memory when Serena MCP is available, else fall back to the plan file (or a workspace state file) so recovery isn't silently broken without Serena.

See @smith-ralph/SKILL.md for full context strategy and retention criteria.

Related

  • @smith-guidance/SKILL.md - Core agent behavior
  • @smith-ctx-claude/SKILL.md - Claude Code: /clear (stop hook enforced)
  • @smith-ctx-cursor/SKILL.md - Cursor: UI indicator, /summarize
  • @smith-ctx-kiro/SKILL.md - Kiro: 80% auto-summarize, Serena memory
  • @smith-serena/SKILL.md - Serena MCP for persistent memory

Before You Finish

Proactive context checks:

  1. Periodically check context (platform-specific method)
  2. At warning threshold: Recommend context reset with retention criteria
  3. At critical threshold: Urgently recommend before degradation
  4. User executes command, agent continues with focused context

Before recommending context reset, prepare:

  • Task requirements summary
  • File paths with line numbers
  • Key design decisions
  • Remaining todos

Recommendation format:

Context at «X»%. Recommend context reset (see platform skill for command).
Keep: «task», «files», «decisions», «todos»
Context Management Skill | Agent Skills