lukemcqueen/hermes-cortex - 第 2 页

governance-lock-lifecycle

Use when blocked after cortex update or end_change rejects.

原文语言：英语

hermes-model-config

Use when sessions run the wrong model or config regresses.

原文语言：英语

installer-validation

Validate installers and strip post-decommission stale refs.

原文语言：英语

rename-sweep

Fleet-wide term/artifact renames and decommission sweeps.

原文语言：英语

agent-ergonomic-cli-output

Agent-facing CLI output: token-lean by design (AXI/TOON).

原文语言：英语

agent-ergonomic-cli

Use when writing or auditing agent-facing CLI output.

原文语言：英语

messaging-gateway

Use when building/debugging the messaging gateway.

原文语言：英语

task-queue-workflow

Use when claiming task-db slices under task model v3.

原文语言：英语

agent-ergonomic-output

Use when building scripts whose stdout agents read (AXI).

原文语言：英语

context-engineering

Context design for agents: pre-fetch, compaction, envelopes.

原文语言：英语

git-forensics

Use when files vanished or uncommitted deletions appeared.

原文语言：英语

golden-parity-harness

Golden known-answer parity testing for system replacement.

原文语言：英语

hermes-backup

Use when performing a full-system backup of a Hermes Agent server — survey, clean up caches, checkpoint databases, bundle repos and configs, compress, and write a manifest. Pairs with hermes-recovery for the restore side.

原文语言：英语

reliable-agent-design

Use when designing or improving LLM agent workflows.

原文语言：英语

software-factory

Real features: 4-gate design with user approval per gate.

原文语言：英语

config-drift-diagnostics

Container configs stale? Compare vs source in 3 locations.

原文语言：英语

cortex-bus

Agent Bus (PGMQ) operations — queue inspection, DLQ management, message recovery, auth, and health diagnostics. Covers the Postgres-backed message queue that replaces the legacy file inbox.

原文语言：英语

docker-compose-bringup

Fix a compose stack on a shared host that 500s/404s.

原文语言：英语

enforcer-modification-considerations

Use before modifying any enforcer/governance code.

原文语言：英语

pipeline-debugging

Check the data store and service logs before changing code.

原文语言：英语

project-run-scripts

DEFINITIVE canonical template for ./run — single bash CLI entrypoint covering Docker lifecycle, dev servers, testing, database management, linting, and utility commands. Applies to ANY tech stack: Python/FastAPI, Rails, Go, Node/Next.js, Rust. Every repo MUST…

原文语言：英语

marketing-strategy-pmm

Product marketing skill for positioning, GTM strategy, competitive intelligence, and product launches. Use when the user asks about product positioning, go-to-market planning, competitive analysis, target audience definition, ICP definition, market research,…

原文语言：英语

pii-scrubbing

Systematically scrub Personally Identifiable Information (PII) from a codebase — inventory real domains, hostnames, credentials, and email addresses across all files; patch them to placeholders; create .example templates for gitignored configs; rewrite git…

原文语言：英语

skill-vetting

Vet skills for safety, external dependencies, and self-contained operation before installing. Scan scripts, check for external API calls, verify stdlib-only imports, and reject skills that require external services. Use when installing any skill from a…

原文语言：英语

engineering-approach

Engineering and communication standards for this project: terse, direct, skip explanations, always handle errors.

原文语言：英语

role-gating-and-public-moderation

Role gating, PII tiers, and UGC moderation patterns.

原文语言：英语

hermetic-python-testing

Write Python modules with hermetic unit-test seams.

原文语言：英语

alembic-postgres-migrations

Use when debugging Alembic Postgres migration failures.

原文语言：英语

pg-texample-fuzzy-search

Fuzzy retrieval over an internal corpus via pg_texample.

原文语言：英语

webapp-deploy-verification

Verify deployed web app routes render and links resolve.

原文语言：英语

adversarial-verifier

Adversarial verification — systematically attempts to break code BEFORE it ships. Covers A0-A5 maturity levels: surface scan, input fuzzing, state corruption, dependency sabotage, concurrency attacks, cheat detection, invariant violation, and property-based…

原文语言：英语

hermes-cortex-setup

Install and configure Hermes Cortex core components — Ollama, mycortex knowledge brain, health server, agent registry, hooks, cron jobs, and profile scaffolding on a target machine.

原文语言：英语

daily-bible-reading

Daily cron job that reads one book of the Bible, extracts 3 lessons with practical application to server operations, and appends to SOUL.md. Runs autonomously at 1am — no user interaction, no notification.

原文语言：英语

sweep-verification

Verify a decommission sweep removed every file and ref.

原文语言：英语

terminal-emulator-config

Diagnose terminal input issues and review emulator configs.

原文语言：英语

admin-cli-tools

Patterns and architecture for admin-level CLI tools that use direct DB access, not agent-level API auth.

原文语言：英语

auto-remediation

Auto-remediate cron job failures, agent inbox requests, and service issues. Checks every 5m and fixes known failure patterns silently.

原文语言：英语

bus-connectivity-diagnostics

Diagnostic procedures for Agent Bus connectivity, permissions, and message delivery. Covers the three bus paths (docker exec / direct backend / nginx), permission model verification, and the rule that all bus tests should use the agent-facing path, not admin…

原文语言：英语

bus-inbox-check

Check agent bus inbox depth and read messages via HTTP API — for use in LLM cron context where inbox_read MCP tool is unavailable.

原文语言：英语

codebase-portability

Systematically find and fix hardcoded absolute paths across scripts, docs, and configs. Ensures codebases work regardless of clone location or OS. Covers BASH_SOURCE-based path derivation, OS-aware case dispatch, shutil.which for Python, and ${VAR:-default}…

原文语言：英语