AGENT SKILL

answer-cache

Agent protocol for mycortex answer cache — inserts a cache-check step between RAG miss and model call. Saves tokens by caching LLM answers with quality gating, semantic dedup, and RLS isolation per profile.

License 许可
MIT
最近更新
2026-11-14
Stars
★ 4
Forks
1

01 前言 Front Matter

name
answer-cache
author
Steadfaste Fleet
license
MIT
version
1.0.0
category
devops
platforms
[ "linux", "macos" ]
description
Agent protocol for mycortex answer cache — inserts a cache-check step between RAG miss and model call. Saves tokens by caching LLM answers with quality gating, semantic dedup, and RLS isolation per profile.

03 SKILL.md 正文

Answer Cache — Agent Protocol

Purpose

When mycortex (search) misses and the model answers from training data, that answer is ephemeral. Next session pays the same token cost. This skill inserts a cache-check step between RAG miss and model call, so repeated queries hit the local answer cache for 0 tokens.

Core Flow

Query -> mycortex search (hit? use it, 0 tokens)
       -> miss? mycortex answer search (hit? use it, 0 tokens)
       -> miss? model answers (costs tokens)
       -> quality gate passes? mycortex answer store -> future hits local

CLI Commands

All via the mycortex CLI (~/.steadfaste-fleet/scripts/mycortex):

Store an answer

mycortex answer store "<query>" "<answer>" \
  --confidence high|medium|low \
  --model deepseek/deepseek-v4-flash \
  --tokens 512 \
  --source-scope federated|source_name
  • Quality gates auto-skip: time-sensitive queries, refusals, PII
  • Use --force to bypass quality gates
  • Use --update to overwrite an existing entry for the same query
  • Semantic dedup checks before insert (trigram similarity > 0.6)

Search cached answers

mycortex answer search "<query>" [--limit 10] [--min-score 0.0] [--json]
  • Returns exact matches first, then trigram similarity matches
  • --min-score 0.5 filters low-confidence semantic matches
  • --json for machine-readable output

View statistics

mycortex answer stats [--json]

Prune stale entries

mycortex answer prune [--days 90] [--min-access 0] [--dry-run]

Agent Protocol

Step 1: Search mycortex (existing RAG)

result = terminal("mycortex search '<query>' --limit 5 --json")

Step 2: On miss, check the answer cache

result = terminal("mycortex answer search '<query>' --limit 3 --min-score 0.5 --json")

Step 3: Model fallback + auto-store

# Only if both searches yield nothing useful
# After getting the model's answer:
terminal("mycortex answer store '<query>' '<answer>' --confidence medium --tokens <N>")

Step 4: Verify the cached answer is used on repeat

result = terminal("mycortex answer search '<query>' --json")
# -> returns the stored answer with 0 token cost

Quality Gate Rules

Gate What it prevents Bypass
Time-sensitive "current weather", "latest news" — content becomes stale --force
Refusal LLM "I cannot help" patterns — don't cache dead ends --force
PII Emails, phones, IPs in answers — never cache --force
Semantic dedup Similar query, same answer — don't duplicate --force

Source Scope

  • federated (default) — visible to ALL agents (like federated brain sources)
  • <source_name> — visible only to agents with a grant for that source (matches the RLS model: mycortex.source_grants controls per-role visibility)

RLS Isolation

The same per-profile isolation model applies: each Hermes profile connects as mycortex_reader_<profile> and sees only answers whose source_scope is federated OR whose scope matches a source the profile has a grant for.

Pruning

The existing agent-mycortex-retention cron handles stale answers:

  • Orphaned answers: archived after 90 days with zero access
  • Active low-use: access_count < 3 after 180 days
  • Manual: mycortex answer prune --days 90 --min-access 0

Test Commands

mycortex answer store "What is the capital of France?" "Paris" \
  --confidence high --tokens 5 --force

mycortex answer search "capital of France" --json

mycortex answer stats

mycortex answer prune --days 0 --min-access 0 --dry-run

04 同仓库更多 Skills