Menú
One-line description of what this fix does and why it matters.
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Aider uses 5min TTL by default and works around long pauses with keepalive pings. Wire up the 1h TTL beta instead.
Aider's --cache-prompts is off by default. Flip it on for supported models.
Stop Cline from burning a cache breakpoint on the volatile current user message every turn.
Cline OpenAI native provider sends no prompt_cache_key. Add a stable per-task key so cached_tokens stops being zero.
Cline's system prompt includes a timestamp that may be recomputed per request, invalidating the system-prompt cache.
Continue's prompt caching is opt-in via config and off by default. Flip the default to systemAndTools.
Basado en la clasificación ocupacional SOC
| name | <skill-slug> |
| description | One-line description of what this fix does and why it matters. |
| target_harness | <harness-name> |
| target_repo | <owner>/<repo> |
| target_files | ["path/to/file.ts"] |
| target_commit | <SHA or branch as of audit> |
| estimated_savings | <e.g. "30-90% input discount on Anthropic" or "0 → 80% hit rate"> |
Which harness, which file, which lines. Cite a permalink to the relevant code at the audit commit.
What's broken in the user's current setup. Be concrete:
usage field comes back zero or wrong--- a/path/to/file.ts
+++ b/path/to/file.ts
@@
- old line
+ new line
If the fix is multi-step or conditional, list the steps.
After applying, the agent should:
<command> to exercise the code path.<specific assertion> — e.g.
usage.cache_read_input_tokens > 0 on the second turn.If the assertion fails, the fix didn't land or there's a second upstream issue. Stop and report.
Why this matters in 1-2 sentences. Link to: