惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - Franky
Apple Machine Learning Research
Apple Machine Learning Research
月光博客
月光博客
Vercel News
Vercel News
Recent Announcements
Recent Announcements
B
Blog RSS Feed
Y
Y Combinator Blog
M
MIT News - Artificial intelligence
MongoDB | Blog
MongoDB | Blog
酷 壳 – CoolShell
酷 壳 – CoolShell
雷峰网
雷峰网
D
Docker
Jina AI
Jina AI
IT之家
IT之家
人人都是产品经理
人人都是产品经理
L
LangChain Blog
G
Google Developers Blog
Google DeepMind News
Google DeepMind News
MyScale Blog
MyScale Blog
博客园 - 叶小钗
The GitHub Blog
The GitHub Blog
The Cloudflare Blog
A
About on SuperTechFans
Hugging Face - Blog
Hugging Face - Blog

Hacker News - Newest: "AI"

AI can't read an investor deck AI as an attorney? Student uses ChatGPT, Gemini to sue UW over alleged racial discrimination Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval GitHub - MeepCastana/KubeezCut: Free Web based video editor Can AI judge journalism? A Thiel-backed startup says yes, even if it risks chilling whistleblowers Coming soon: 10 Things That Matter in AI Right Now DARPA built an AI to fact-check enemy weapons claims What explains heterogeneity in AI adoption? When AI Meets Muscle: Context-Aware Electrical Stimulation Promises a New Way to Guide Human Movements - Department of Computer Science AI Changed How We Build. It Did Not Change What Matters. Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Meta spins up AI version of Mark Zuckerberg to engage with employees Code Mode: Let Your AI Write Programs, Not Just Call Tools | TanStack Blog GitHub - Delavalom/graft: Go framework for building AI agents. Type-safe tools, multi-provider (OpenAI, Anthropic, Gemini, Bedrock), zero vendor SDKs. India's TCS tops estimates, says new AI models did not dent services demand Gen Z's fading AI hype Strong feeling: we are in a folded AI reality GitHub - machinarii/total-recall-catalog: A reference catalog of latest knowledge retrieval, memory & RAG systems GitHub - mensfeld/code-on-incus: Give each AI agent its own isolated machine with root, Docker, and systemd. Active defense detects and stops threats automatically.. Quantization, LoRA, and the 8% Problem: Benchmarking Local LLMs for Production AI Iran war: We spoke to the man making Lego-style AI videos that experts say are powerful propaganda Powell, Bessent discussed Anthropic's Mythos AI cyber threat with major U.S. banks GitHub - immartian/bellamem: Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact. recursive-mode: The Repo-Native Operating System for AI Engineering After the attack on Sam Altman's home, will AI CEO's go on the offensive? The biggest advance in AI since the LLM Opus 4.6 vs GPT 5.4 One Prompt Unity World Generation Test “AI polls” are fake polls Client Challenge Can AI be a 'child of God'? Inside Anthropic's meeting with Christian leaders
GitHub - ibm777p2/Intelligence-Emotions: Intelligence Emo...
ibm777p2 · 2026-06-13 · via Hacker News - Newest: "AI"

Your mind is your best friend — and, left unsupervised, your worst critic.

Intelligence Emotions turns Claude Code into your personal mental-fitness team, built on the Intelligence Emotions model. Five coaches, one practice: catch the ten saboteurs (led by the Judge — the inner voice that finds fault with everything), do PQ reps (10-second shifts of attention to a physical sensation), and let the Sage choose the response. The team remembers between sessions — a private, local-only journal that makes every coach smarter about your actual patterns.

This is not therapy or medical care. The coaches say so themselves, once, plainly — and they know when to stop coaching and point you to a professional instead.

Who this is for:

  • Anyone building mental fitness who wants the practice to actually happen daily
  • First-time practitioners — structured coaching instead of a blank journal
  • People with a challenge they keep circling — a habit, a conflict, a decision that won't resolve

Quick start

  1. Install Intelligence Emotions (30 seconds — see below)
  2. Run /sage-session — bring one real challenge
  3. Run /intercept the next time a thought stings
  4. Run /pq-retro tonight before closing the laptop
  5. Run /pq-score after three days of journal data
  6. Stop there. You'll know if this is for you.

Install — 30 seconds

Requirements: Claude Code, Git, Bun v1.0+

Step 1: Install on your machine

Open Claude Code and paste this. Claude does the rest.

Install Intelligence Emotions: run git clone --single-branch --depth 1 https://github.com/YOUR-GITHUB-USERNAME/intelligence-emotions.git ~/.claude/skills/pq && cd ~/.claude/skills/pq && ./setup then add an "Intelligence Emotions" section to CLAUDE.md that lists the available skills: /pq, /sage-session, /intercept, /saboteur-scan, /sage-perspective, /navigate-review, /pq-retro, /daily-pipeline, /growth-spec, /commit, /pq-score, /habit-watch, /life-design, /saboteur-audit, /insight-doc, /second-coach, /context-save, /context-restore — and says that /pq is the front door when unsure which to use.

(The skills use the pq prefix — your mental-fitness quotient, the percentage of time your mind serves you rather than sabotages you.)

Step 2: First session

Bring something real — a challenge with weight, however half-formed. The Sage takes it from there.

See it work

You:  /intercept my coworker got promoted and I feel worthless

Spotter:  That word "worthless" — let's trace where it came from.
          What's the actual sentence running in your head, word for word?

You:  "she's two years younger and already ahead of me"

Spotter:  Two patterns in that sentence. The comparison clock — that's the
          Hyper-Achiever (the pattern that pegs your worth to external
          scorekeeping). And "worthless" is the Judge delivering the verdict.
          The lie: her promotion is a measurement of you. It feels like data.
          It's machinery — the same machinery would have produced the same
          sentence about whoever got promoted.

          Next time a promotion email lands — anyone's — ten seconds on your
          feet on the floor before your mind gets the microphone.
          Want me to log this interception?

The practice

The daily loop, as a team:

Time Command Coach What happens
Morning /daily-pipeline The Trainer One intention, a saboteur forecast for the day's terrain, first rep
In the moment /intercept The Spotter Trace a stinging thought: trigger → saboteur → the lie → the need. Rep prescribed, interception logged
Evening /pq-retro The Witness Five minutes: what triggered, what got caught, rep count, one Sage win. Writes the journal
Anytime /sage-session The Sage The flagship — any life challenge, walked through all five Sage powers

The 21-day arc, when something deserves real practice:

Step Command What happens
1 /growth-spec A vague intention ("be more patient with my kids") becomes a concrete 21-day practice: trigger inventory, daily reps anchored to real moments, check-ins
2 /habit-watch Scheduled look-ins; catches drift while the fix is still one sentence
3 /commit Day 21: the honest review, the closing entry, the commitment marked. No grades — what happened, what it taught, what's next

Which coach should I ask?

Situation Command
A whole situation with weight /sage-session
One specific moment that stung /intercept
"Check this draft/entry before I send it" — no fixes wanted /saboteur-scan
A decision I keep flip-flopping on /sage-perspective
"Will this plan actually stick?" /navigate-review
"Map all my patterns" — the wide view /saboteur-audit
"How am I doing?" — numbers from real data /pq-score
Design my mornings / environment from scratch /life-design
Something just clicked — keep it /insight-doc
"Push back on me harder" /second-coach
Not sure /pq

Power tools

Your data, on the command line (all local, all yours):

~/.claude/skills/pq/bin/pq-journal-search --days 7           # this week's interceptions, entries, commitments
~/.claude/skills/pq/bin/pq-journal-search --stats --days 7   # per-day reps, interception rate, saboteur tallies
~/.claude/skills/pq/bin/pq-journal-log saboteurs '{...}'     # log an interception by hand
~/.claude/skills/pq/bin/pq-journal-log entries --redact <id> # forget a record — expunged completely, archive included
~/.claude/skills/pq/bin/pq-config list                       # your preferences (name, rep_target, cadence, explain_level)

The journal is event-sourced and append-only: three streams under ~/.pq/journal/ (entries.jsonl, saboteurs.jsonl, commitments.jsonl), supersede-don't-delete, redact-on-demand. A write-time guard refuses credential-shaped secrets. /pq-score and /saboteur-audit read this data — nothing is ever estimated.

Uninstall

Option 1: Run the uninstall script

~/.claude/skills/pq/setup --uninstall

Removes the skill symlinks. Your journal and config in ~/.pq are never touched — that data is yours, including after you leave.

Option 2: Manual removal

# 1. Remove the skills
rm -rf ~/.claude/skills/pq
for s in pq sage-session intercept saboteur-scan sage-perspective navigate-review \
         pq-retro daily-pipeline growth-spec commit pq-score habit-watch life-design \
         saboteur-audit insight-doc second-coach context-save context-restore; do
  [ -L ~/.claude/skills/$s ] && rm ~/.claude/skills/$s
done

# 2. Only if you also want your journal gone (irreversible — it's your history):
# rm -rf ~/.pq

Docs

  • CLAUDE.md — contributor guide: the pipeline, the iron rules, how to write a skill
  • ETHOS.md — the operating philosophy: label don't fight; ten seconds at a time; every setback is a gift; the coach is never the Judge
  • CHANGELOG.md — versions, starting fresh at 1.0.0.0

Privacy

There is no telemetry. None, not opt-out — none exists. Everything personal lives in ~/.pq/ with owner-only permissions (0700 directories, 0600 files), never pushed to any remote, never sent to any service. Redaction is honored absolutely: ask any coach to forget something and the record is expunged from the live log and the archive. The test suite enforces that no external sink can quietly appear.

Troubleshooting

  • Skills don't show up in Claude Code — re-run ~/.claude/skills/pq/setup; check that ~/.claude/skills/sage-session is a symlink pointing into the install.
  • bun: command not found — install Bun, then re-run ./setup.
  • A coach sounds like the Judge — that's a bug in the product, not a feature of honesty. Open an issue with the transcript line; the anti-Judge rule is the product.

Intelligence Emotions

Exactly 10 saboteurs. Exactly 5 Sage powers. Never an invented eleventh or sixth — the test suite enforces the canon. The team: the Sage (lead coach), the Spotter (pattern recognition), the Trainer (rep drills), the Navigator (plans and direction), the Witness (reflection). One rule outranks everything they say: no coach is ever the Judge.

License

MIT — see LICENSE.