Skip to content

prangishviliAbe/agent-skills / token-efficiency

token-efficiency

Verified

Make responses and agent work cheaper to read and to run without losing correctness: concise answers, TL;DR, brief status updates and final summaries, just-the-code replies, and lean tool use that avoids rereading files, dumping logs, and repeating checks. Use whenever the user asks to be brief, short, or concise, says "less text", "just the code", "TLDR", or "save tokens", complains about limits or cost, or runs a long session where context is filling up. Puts the answer first, keeps the evidence, failures, and caveats that change decisions, replies in the user's language, and never trades correctness, completeness of the work, or quality of the result for brevity.

FileKindSize
SKILL.mdskill md5.8 KB
agents/openai.yamlconfig286 B
references/context-economy.mdmarkdown1.5 KB
references/output-craft.mdmarkdown1.9 KB

references/context-economy.md

View on GitHub
# Bounded investigation and useful context

Read when: Inspecting large logs/files, managing long work, or preparing a handoff.

## Search before reading

```bash partial
rg -n -i 'error|exception|failed' work/ci.log
rg -n 'suspectedSymbol' src
rg --files src
```

Read bounded context around a relevant match, then trace upstream to the first causal failure. A final CI summary can hide the original exception. Avoid printing secrets or entire logs.

| Input | Efficient move | Preserve |
| --- | --- | --- |
| Large log | Search, inspect bounded ranges, follow first cause | Exact location and surrounding evidence |
| Large repository | Find entrypoint/callers, read the relevant path | Constraints and dependency relationships |
| Repeated progress | Report meaningful state changes only | Blocker or next unresolved decision |
| Long task | Maintain a compact state note | Objective, constraints, edits, evidence, next action |
| Broad comparison | Read only sources that can change the choice | Reasons and material uncertainty |

Do not invent universal claims about token pricing or latency. Caching, context reuse, model behavior, and billing vary by runtime.

## Handoff artifact

```text
Objective:
User constraints and authorization:
Current artifact/revision:
Completed changes:
Verified checks:
Inferred/Assumed:
Not run and why:
Next required action:
```

A good handoff prevents redoing work and preserves why a decision was made. Retain abandoned approaches only when they explain a surviving tradeoff.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall