prompt-evaluation-runn…
Use when evaluating prompts, LLM outputs, red-team suites, or model behavior with local eval configs and safe provider/cost controls.
Use when deciding whether to compact agent context -- the window is filling up, you are about to switch from one phase to another (research, planning, implementation, debugging), a failed approach left dead-end reasoning, or a previous mid-task compaction lost critical state.
$ npx -y skills add yeaight7/agent-powerups --skill strategic-context-compaction --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/strategic-context-compactionContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when deciding whether to compact agent context -- the window is filling up, you are about to switch from one phase to another (research, planning, implementation, debugging), a failed approach left dead-end reasoning, or a previous mid-task compaction lost critical state.
name: strategic-context-compaction description: Use when deciding whether to compact agent context -- the window is filling up, you are about to switch from one phase to another (research, planning, implementation, debugging), a failed approach left dead-end reasoning, or a previous mid-task compaction lost critical state.
Compact at logical boundaries to preserve high-value context while clearing noise. Arbitrary or mid-task compaction loses critical state -- file paths, variable names, partial reasoning -- that is expensive to reconstruct.
1. **Decide whether to compact by transition.** Use the boundary table; default to NOT compacting mid-work.
| Transition | Compact? | Reason | | --- | --- | --- | | Research -- Planning | Yes | Research context is bulky; the plan is the distilled output | | Planning -- Implementation | Yes | Plan is saved in tasks/files; context is free to reset | | Implementation -- Testing | Maybe | Keep if tests reference recent code; compact if switching focus area | | Debugging -- Next feature | Yes | Debug traces pollute unrelated work | | Mid-implementation | No | Losing file paths, variable names, partial state is costly | | After a failed approach | Yes | Clear dead-end reasoning before trying a new approach |
2. **Save anything you cannot reconstruct cheaply before compacting.**
3. **Know what survives versus what is lost.** Anything in the Lost column must be persisted in step 2 first.
| Survives | Lost | | --- | --- | | CLAUDE.md / AGENTS.md instructions | Intermediate reasoning | | Task list (TodoWrite) | File contents read in session | | Files on disk | Tool call history | | Git state | Verbally stated preferences | | Memory files | Multi-step conversation context |
4. **Compact with an intent-carrying summary prompt, then restore.**
Curated power-ups for coding agents: skills, slash commands, MCP configs, hooks, AGENTS.md templates, and workflows for serious software engineering. Claude Code, Codex, Antigravity CLI, Cursor and more
Repo: yeaight7/agent-powerups
Use when evaluating prompts, LLM outputs, red-team suites, or model behavior with local eval configs and safe provider/cost controls.
Use when creating or reviewing red-team eval plugins, attack templates, grader rubrics, safety fixtures, or model-risk test metadata.
Use when designing, running, debugging, or hardening deterministic eval suites for agent skills, prompts, tool workflows, or MCP-backed cases.
Use when designing tool definitions for a new agent or subagent, an agent shows high retry rates, ambiguous tool invocations, or silent failures, or an…
Use when routing a prompt to a local provider CLI for a second opinion, review, or plan -- you are about to call a provider directly, need the response saved…
Use when starting work in an unfamiliar area of a codebase, spawning a subagent that needs targeted file context, a first search pass missed the relevant file,…