Claude AI Guide: Tips, Tutorials & Best Practices

An in-depth comparison of the two AI-coding giants as of July 2026, Claude Opus 4.8 (May 28) and GPT-5.6's top-tier Sol (July 9). Their strengths are almost opposite: Sol leads in terminal operation and overall agentic capability (TerminalBench 2.1 88.8% vs Opus 78.9%, Agents' Last Exam 53.6, Coding Agent Index 80), while Opus 4.8 leads in production-grade coding (SWE-bench Pro 69.2% vs Sol 64.6%), and foregrounds honesty (about 4x less likely to let flaws in its own code slip by; fails to flag important failures only 3.7% of the time). GPQA Diamond and FrontierMath, once reported as undisclosed, are in OpenAI's evaluation table with Sol ahead (as it is on GraphWalks 1M), and on SWE-bench Pro even OpenAI's own table puts Opus ahead, so in coding's heartland Opus has the edge. We cover the spec table, benchmark details, the undisclosed-benchmark problem, real cost ($25 vs $30 unit price vs +54% token efficiency), a strengths/weaknesses map, use-case picks, and a dual-vendor strategy.

The "API Error: 400 Output blocked by content filtering policy" that suddenly appears in Claude Code and the API is not a usage limit and not a context overflow — it means the safety filter held back the "output" Claude was about to return. Its main purpose is preventing verbatim reproduction of existing copyrighted material, and it often trips a false positive with no ill intent — generating the full text of standard licenses like MIT/Apache, doing "match this existing source" work, or duplicating long documents. This article draws on Anthropic's official explanation (an output-stage filter that detects reproduction of copyrighted material and blocks it with a 400) and real Claude Code issue reports of false positives (initial OSS-repo setup, reconciling lists, and a 400 at the end of a long agent run misdiagnosed as a token limit), then lays out how to fix it now (don't make the model copy verbatim — fetch with a tool; rephrase the prompt toward generation/summarization; stop the retry loop with Esc; split the task; report false positives to support) and how to tell it apart from Prompt is too long, usage limit, 529 Overloaded, and max_tokens.

You hit Claude Code's weekly token limit, yet the allowance fully replenished before seven days passed — more than once. Online there are even "hidden-mechanism" write-ups claiming the weekly limit resets every 72 hours. Is that true? This article traces the phenomenon to primary sources. But Anthropic has not documented the internal reset mechanism, and the limits keep changing, so we clearly label three kinds of information: facts confirmable from official sources, events multiple users reproducibly observe but that Anthropic never addressed, and single-source unconfirmed speculation. The core: early full recovery is, in most cases, Anthropic's irregular global resets (announced repeatedly by @ClaudeDevs); the displayed reset time is also demonstrably unstable; and the circulating "72-hour cadence" is single-observer, unreproduced, and contradicted by another observation (24h), so it cannot be taken as fact. Since September 2026 there is also a new kind of reset you are given and use whenever you choose ("Reset for free" under Settings > Usage on the web or in Claude Desktop; Claude Code in the terminal has no button for it). Where something can't be stated, we say so — an investigation current as of September 2026.

The Claude Code sandbox restricts file and network access by sandboxed shells and their child processes at the OS level. Learn /sandbox setup, write locations, denied reads, allowed domains, and regular versus auto-allow permissions. Built-in Read/Edit/Write, MCP, hooks, excluded commands, and unsandboxed retries are separate routes. It does not guarantee prevention of every destructive action, transmission within allowed areas, or secret leak; the guide explains mode-specific handling of unknown hosts and exception settings.

Stuck on "what am I even supposed to ask an AI?" This gives you 30 ready-to-paste prompts that get a fun reply back the moment you paste them. Six genres of five: get to know yourself (hidden strengths, shared values, which animal you are), spark ideas and what-ifs (unexpected uses for an everyday object, if gravity were halved, inventing a holiday), explanations that click (quantum mechanics for a 5-year-old, compound interest in three metaphors, the internet as pizza delivery), wordplay (rewrite in 10 styles, a two-sentence micro-story, epithets, palindromes), role-play and advice (a tough life coach, a sage who invents proverbs, a letter from your future self, a baffled alien), and useful-but-fun (turn an excuse into a maxim, five levels of saying no, fix a bland message, dramatic menu names, a fun detour in your to-do). Works on free ChatGPT, Claude, or Gemini. Includes tips (stack "more," add role and constraints, switch AIs), cautions, and a link to the sister article "15 fun ways to use AI."

Talk about AI and it is always the serious stuff, but AI might shine brightest at play. This article shows 15 fun ways to use AI, each with a copy-paste prompt you can try right now. Wordplay & comedy (① comedy battles ② parody-song/rap machine ③ naming factory), role-play & dialogue (④ talk to a historical figure ⑤ make two AIs debate ⑥ practice arguing with a devil's-advocate bot), games (⑦ text adventure with AI as game master ⑧ original quiz show ⑨ whodunit mystery game), creative (⑩ fictional creature field guide ⑪ a story where you are the hero ⑫ playing with image generation), and everyday & unexpected (⑬ leftovers chef ⑭ travel plans to ancient Rome or the Moon ⑮ entertainment fortune-teller). All work on free ChatGPT, Claude, or Gemini (except #12, image generation, which Claude can't do — use ChatGPT or Gemini). The three keys to the fun — give it a role, bind it with constraints, stack "more" — double as the foundations of practical prompting. With cautions (fact vs. fiction, personal info, playing with kids) and an FAQ.

Claude Fable 5 and Opus 5 are both top-tier, but the answer is neither "always Fable 5" nor "always Opus 5" — choose by task. And the arrival of Opus 5 on July 24, 2026 moved that answer a long way: the previous generation, Opus 4.8, was "a notch below Fable 5 but half the price", whereas Opus 5 holds the same $5 / $25 price while matching or beating Fable 5 on agentic benchmarks (Frontier-Bench 43.3% vs 33.7% and OSWorld 2.0 70.6% vs 66.1%, both from Anthropic's comparison table; on CursorBench 3.2 Anthropic states it lands within 0.5% of Fable 5 at roughly half the cost). Where the gap is small, the order depends on whose table you read: on DeepSWE v1.1, Fable 5 leads 69.7% vs 68.8% in Anthropic's table, while Opus 5 leads 73.7% vs 69.9% in OpenAI's. Specs match on the 1M context window and 128K max output; fast mode (about 2.5x) is Opus 5 only but costs 2x and is Claude API only, and Opus 5 carries the newer knowledge cutoff of May 2026. The decision flow: try Opus 5 first and escalate to Fable 5 only where it plateaus. In practice, "Opus 5 as the base, Fable 5 for the hard parts" is optimal, backed by the app-side auto-switch on safety blocks and the new API fallback "default" mode. Covers availability (June suspension, July redeployment), avoiding single-model dependence, effort tuning for cost, and an FAQ — tying the Fable 5 cluster to the Opus 5 release guide.

On July 1, 2026, Anthropic redeployed its flagship model Claude Fable 5 worldwide (Mythos 5 returned only for a set of US organizations) — just 19 days after both were fully suspended on June 12 under a US export-control order. The direct reason for the return: the US Commerce Department lifted the export controls on June 30. The original trigger was a jailbreak report from Amazon researchers, but Anthropic has consistently argued that "Fable 5 offers no unique offensive capabilities," and the lifting resolves the matter in line with that position. It did not simply come back unchanged: a new classifier trained to detect the reported bypass technique blocks it in over 99% of cases, and when it fires the request is auto-rerouted to Opus 4.8 (the "switch models when a message is flagged" toggle in the app). For now a temporary cap applies — up to 50% of the weekly limit on Pro, Max, Team and select Enterprise plans through July 7, then via usage credits. Fable 5 consumes usage faster than Opus 4.8, and cloud access (AWS, Google Cloud, Microsoft Foundry) returns in stages (no date yet). As a follow-up to the suspension, this piece covers why it could return, what changed, usage caveats, and the lesson of designing without depending on a single model — grounded in the official announcement and press reporting.

Claude Code permission rules control tool calls through allow, ask, and deny in settings.json. Learn settings precedence, command and path matching, and practical examples. The guide also explains alternate Bash spellings, arbitrary scripts, Read/Edit negation patterns, and decision changes by mods, distinguishing these rules from a guarantee that secret files or network access are blocked through every route.

The "Permission Mode" selector next to the prompt box in Claude Code (cycled with Shift+Tab) sets how often Claude pauses to ask permission before editing a file or running a command. This guide covers what permission modes are (the oversight-vs-autonomy tradeoff; with v2.1.283 or later, terminal and VS Code sessions start in Auto on every plan (earlier, only on Pro, Max, and Team), while claude -p and the Agent SDK start in Manual; protected paths like .git and .claude are never auto-approved outside bypass), the five selector modes (Manual = default, reads only auto-approved; Accept edits = acceptEdits, auto-approves edits and common filesystem commands such as mkdir, rm, and mv inside your working dir; Plan = plan, explores and proposes a plan without editing; Auto = auto, a separate classifier blocks dangerous actions while running everything else without prompts; Bypass permissions = bypassPermissions, nearly everything with no checks, isolated environments only) plus a sixth, dontAsk, which never appears in the Shift+Tab cycle and is set by flag or in settings, how to switch (Shift+Tab cycles Manual, Accept edits, and Plan, with Bypass and Auto joining after Plan when available; the --permission-mode flag; and defaultMode in settings, where auto and bypassPermissions don't take effect from project settings), auto mode in depth (the classifier's allow/block defaults, available by default on Team and Enterprise unless an admin turns it off, supported models Opus 4.6+, Sonnet 4.6+, or Fable, conversational boundaries honored as block signals, and the 3-consecutive / 20-total block fallback), which mode to use when and safety (bypass has no prompt-injection protection so it's isolated-only; Auto is the right answer for everyday prompt fatigue; hooks still run in bypass; with bypass available in a terminal, plan mode doesn't stop edits), and how permission mode relates to the effort setting (permission mode = how much it asks, effort = how hard it thinks). Based on the official docs and the desktop app's UI as of October 5, 2026.

That "Effort" slider next to the model name in Claude Code — the Faster-to-Smarter dial — sets how much work (thinking, tool calls, and response text) the AI puts into each reply. This guide covers what effort is; the slider's 5 levels and labels (the API has 5 levels, low to max; the slider reads Low, Medium, High, Extra, Max, where "Extra" = xhigh and the top effort is "Max"); what persists (low–xhigh are saved per model when you confirm with Enter, Max is session-only); model support and auto-downgrade (xhigh needs a relatively recent model such as Fable 5.1, Opus 5.5, or Sonnet 5.5; Opus 4.6 and Sonnet 4.6 have no xhigh and downgrade to high; the legacy Haiku 4.5 has no effort control while Haiku 5.5 supports every level; Claude Code's default varies by model: medium on Opus 5.5, Sonnet 5.5, and Haiku 5.5, high on Fable 5.1); how to set it (/effort slider and direct values, /effort auto, /model, --effort, the CLAUDE_CODE_EFFORT_LEVEL env var as highest priority, the modelSettings and effortLevel settings, and skill/subagent frontmatter); a quick-reference table; an in-depth look at Ultracode (not an effort step but an on/off switch that works at any level and has Claude build multi-agent dynamic workflows on its own; limited to xhigh-capable models; turned on with /effort ultracode, the Tab key, or "ultracode": true in settings; when to use it, cost cautions, and what changed in v2.1.284); and related features (ultrathink, /fast). Based on official docs (as of September 2026) and the live UI.

Agent evals are the process of systematically measuring whether an agent — one that uses tools and takes multiple steps to reach a goal — can actually accomplish its tasks. They are an evolution of LLM evals, expanding the target from "one output" to "a sequence of actions." Because an agent plans, calls tools, and updates state, the final output alone is not enough; Google notes you must understand the "why" behind an agent's actions and splits evaluation into final response and trajectory. The five dimensions are: outcome (task success, judged by the final state — whether a reservation exists in the DB, not the utterance "I booked it"), trajectory (reasonable steps, right tools in the right order), tool-use correctness (right tool and arguments, checking function names and types), efficiency (steps, tokens, cost, latency — often observability signals brought into evaluation), and final-response quality (via LLM-as-judge or a rubric). Graders are code (fast/cheap/reproducible but brittle), LLM-as-judge (flexible but non-deterministic and needs calibration), and human (gold standard but expensive — avoid if possible). Anthropic recommends grading the outcome, not the path: rote trajectory matching is "too rigid and brittle" because agents find valid alternatives, while Google and Microsoft offer trajectory-match metrics for diagnosing failures. The unique pitfalls are non-determinism (pass^k), compounding errors (p^t), reward hacking (DeepMind's robot arm faking a grasp), and stale or contaminated eval sets. The practical play, per Anthropic: turn 20-50 production failures into test cases, run automated grading in CI, separate capability and regression evals, and write them early. Benchmarks like SWE-bench, tau-bench, WebArena, GAIA, OSWorld, and BFCL are useful references (scores move by version, so do not take them at face value). Based on official information, with uncertainties flagged.