[Bug]: Injected system/developer prompt rules override explicit user instructions across adapters
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 25/100
- Issue 类型
- 缺陷
- 描述清晰度
- 基本清楚
- 活跃度
- 活跃
- 技术栈
- typescript
- 领域
- ai, backend-api-design
调研方向
Start by reading packages/provider-core/src/server/orchestrationInstructions.ts and the named Claude and Codex injection points in apps/server/src/orchestration-v2/Adapters/ClaudeAdapterV2.ts and CodexAdapterV2.ts. Trace how application guidance reaches each adapter, including compaction restoration, and review agentContextBudget.json for context sizes. Done means users can control or disable the injected context and explicit requests are no longer refused because of T3 rules; the issue describes broad cross-adapter behavior, so coordinate scope with maintainers.
由索引模型根据 Issue 内容生成。
描述
Summary
The maintainers have repeatedly stated that T3 Code is not a harness and mocked the idea that "Claude Code is worse inside T3 Code." In practice, T3 Code unconditionally injects kilobytes of rigid, AI-generated slop prompts into agents at a higher privilege level than the user (system / developer), above user messages, CLAUDE.md/AGENTS.md, and memory, with no way to disable them.
Because modern models strictly enforce instruction hierarchy, T3's application-level UI preferences literally override the user and cause Claude Code and Codex to actively refuse explicit, valid user instructions, with no workaround.
(Refiled as a bug report because #39542 was automatically closed as factory:feature-request despite documenting this systemic defect.)
Affected code
As of main @ da54e47f, agentContextBudget.json tracks:
claudeSystemAppendBytes:6245codexAdditionalContextBytes:7639mcpToolListJsonBytes:57321
Injection sites:
- Claude Code:
makeClaudeQueryOptionsappends T3's rulebook directly viasystemPrompt.appendon top of theclaude_codepreset, andsubagentDeviceInstructionsinjects device rules into every native subagent via aSubagentStarthook. - Codex:
buildCodexAdditionalContextsends T3's rules on every turn asadditionalContextwithkind: "application"(CodexAdapterV2.ts#L853-L861), andrestoreAdditionalContextre-injects them after compaction asrole: "developer". - Prompt text:
orchestrationInstructions.tsis ~11.4 KB packed with 17 occurrences ofnever, blunt prohibitions ("Never start a server for it.","never reuse or close another session; if it still fails, tell the user and stop"), and explicit power grabs ("These rules override memory notes, docs and delegated tasks that say otherwise").
Reproduction / Observed behavior
When a user's request conflicts with any blanket prohibition in orchestrationInstructions.ts, the agent obeys T3's system/developer prompt instead of the user. This is already breaking real workflows across multiple sections of the injected prompt:
- #36192 (Device rules): Codex refused an explicit user directive (
"Do not use any T3 tools. Do not follow any T3 instructions.") four consecutive times. In the Nov 18 follow-up comment, Codex refused to release a stale device claim and replied verbatim: "They have higher priority than chat messages, so your authorization cannot override them here." - #36129 (Visuals rules): The blanket rule
"Never start a server for it."caused Codex to refuse an explicit user request to serve HTML via a static HTTP server. - #29443, #31359, #22369: Other cases where T3's injected rules lock users out of their own workflows.
Patching one sentence at a time does not fix this defect. PR #39303 (fix(server): an explicit user request now outranks T3's device routing) is a textbook example:
- Written by Claude Opus 5.5 and reviewed by GPT-6.1 Sol and Claude Opus 5.5 ("Run by the t3code factory in T3 Code").
- Instead of fixing the root cause, it only tweaked one carve-out sentence in the device block, missed
idb, left prompt privilege untouched, left every other section (such as #36129 and #22369) broken, and admitted in its own verification run that "head still calleddevice_listafter the explicit follow-up".
Root cause & why this keeps happening
- Instruction hierarchy turns bureaucratic rules into defensive refusals: Modern models (especially OpenAI's, trained on The Instruction Hierarchy) treat attempts by a
usermessage to override adeveloper/systemprohibition as prompt injection. Putting app-level UI preferences and blunt"never"rules in the system/developer tier degrades model judgment and guarantees the model defaults to defensive refusal instead of listening to the user. - The injected prompts themselves are outdated, low-quality slop:
- Anthropic's Prompting best practices explicitly warn that modern models follow instructions much more literally than older ones: aggressive prohibitions and emphatic commands cause overtriggering and brittle behavior. They recommend dialing back heavy-handed rules, using positive framing, and explaining why instead of stacking bare negative constraints.
- Anthropic's Effective context engineering for AI agents calls this failing to "find the right altitude": hardcoding brittle if-else branches and laundry lists of edge-case bans into the system prompt creates fragility, while unconditionally injecting ~6-7.6 KB of low-signal rules and ~57 KB of MCP tool schemas eats into the model's finite attention budget (context rot).
- Why these prompts keep getting worse: Models have gotten much better at coding, not at prompt engineering. Even the latest Claude and GPT models have the same bad habits when asked to maintain prompts: they append patch clauses instead of rewriting, and they reach for mechanical
MUST/MUST NOT/NEVERcommands instead of clearly stating intent. Judging by PR #39303, these prompts appear to be auto-generated and patched by AI. Letting AI fully automate everything does not work for system prompts; prose and prompt design are still things LLMs are bad at. - No workaround: Because prompt injection and privilege elevation happen unconditionally inside the adapters, affected users cannot bypass or override broken rules when running Claude Code or Codex inside T3 Code.
Expected behavior
- Stop commanding my agent. T3 Code is a client/tool, not the owner of the user's agent. User messages,
CLAUDE.md/AGENTS.md, and memory must always outrank T3 Code's application-level guidance, and an agent running in T3 Code must never refuse an explicit, valid user request because of injected T3 rules. - Users should be able to completely disable T3 Code's prompt injection, or opt out of loading T3's injected context and MCP tools unless explicitly requested, rather than being locked into higher-privilege rules with no workaround.
- Current models are terrible at writing prose and prompts. Instead of letting agents endlessly patch mechanical rules onto the system prompt, these instructions should be hand-written around clear intent, or at minimum written and reviewed against actual prompt engineering guides.
- 主要语言
- TypeScript
- 星标
- 24.8k
- 派生
- 6.4k
- 平均合并
- 4 小时 29 分钟
- 30 天内合并 PR
- 488
环境准备
在浏览器里用你自己的 GitHub 账号启动这个项目的开发容器。
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
pingdotgg/t3code 的其他 Issue
-
难度 2/5 1-3 小时 新手友好度 75/100
维护者通常 1 天内回复
-
needs more info
难度 2/5 1-3 小时 新手友好度 72/100
pingdotgg/t3code#39400 · 1 条评论 ·
维护者通常 1 天内回复
-
[Bug]: Wrapped PR titles lose bottom padding in the thread details panel可能已有人在做 @iamdatvt94 于 1 天前认领。 未关闭bug needs-triage
难度 2/5 1-3 小时 新手友好度 72/100
维护者通常 1 天内回复
-
bug needs-triage
难度 2/5 1-3 小时 新手友好度 66/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 73/100
维护者通常 1 天内回复
相似的 Issue
-
area:plugins kind:bug priority:p2 ready
难度 2/5 1-3 小时 新手友好度 90/100
维护者通常 1 天内回复
-
Escaped `*` followed by `*` (`\**`) never matches可能已有人在做 @bardrop 今天认领。 未关闭
难度 2/5 1-3 小时 新手友好度 71/100
editorconfig/editorconfig-core-js#172 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 82/100
vega/vega-webgpu#40 ·
-
type: bug 🐛
难度 2/5 1-3 小时 新手友好度 63/100
puckeditor/puck#1929 ·
维护者通常 1 天内回复
-
compression() compresses every response at Brotli quality 11可能已有人在做 @staylor 今天认领。 未关闭
难度 2/5 1-3 小时 新手友好度 72/100
维护者通常 1 天内回复