Mert Koseoglu
d05a37584e
Issue #482 (makoMakoGo) reported that context-mode's caveman/terse injection pressures the model toward brevity on its FINAL ANSWER, not just on tool-output reporting. Cited evidence: Moonshot AI on kimi-k2.5 (anomalyco/opencode#20258, PR #20259) — aggressive brevity prompts measurably degrade coding/reasoning benchmarks because the model drops assumptions, caveats, verification evidence, failure modes, and security warnings the user actually needs. Considered: A — config switch ("injectCommunicationStyle: false"). Rejected: switches default-on become dead code. B — close FR with rationale. Rejected: ignores valid evidence. C — refine wording ("compress when reporting raw tool output, be complete for technical answers"). Rejected: still text injection, model-dependent, half-measure. D — full strip everywhere. Adopted. The decision after grilling: context-mode's value is data routing (sandbox, FTS5, session continuity), not prose styling. The brevity injection conflated three goals — keeping raw data out of context (real, hard-enforced), summarizing tool output compactly (LLMs auto- calibrate), and final-answer prose style (the wrong target). Strip all 22 sites where prose-style language landed. Sites stripped (A-Z): hooks/routing-block.mjs - <communication_style> block (Terse like caveman, fragments OK, auto-expand for security warnings) - <response_format> block (Concise summary, 2-3 bullets) src/server.ts (5 MCP tool descriptions + 2 cosmetic comments) - ctx_execute "When reporting results — terse..." - ctx_execute_file same - ctx_search same - ctx_fetch_and_index same (URL + commands shapes, both) - cosmetic comment "Caveman style — terse status line" - rewrote concurrency note: "Indexing is serial regardless of concurrency" → "Fetches parallelize up to your concurrency setting; FTS5 indexing serializes the writes after (SQLite single-writer rule)." — same fact, less jargon. configs/ (15 adapter MD files — every shipped system prompt) antigravity/GEMINI.md, claude-code/CLAUDE.md, codex/AGENTS.md, cursor/context-mode.mdc, gemini-cli/GEMINI.md, jetbrains-copilot/ copilot-instructions.md, kilo/AGENTS.md, kiro/KIRO.md, omp/SYSTEM.md, openclaw/AGENTS.md, opencode/AGENTS.md, pi/AGENTS.md, qwen-code/ QWEN.md, vscode-copilot/copilot-instructions.md, zed/AGENTS.md All had identical "## Output" block: 3 caveman lines stripped, workflow lines ("Write artifacts to FILES", "Descriptive source labels") kept. CLAUDE.md (repo root — internal dev instructions) Same caveman block stripped. We don't ship this file but we do eat our own dog food. README.md Pillar 4 ("Output Compression — Terse like caveman...") rewritten to "No prose-style enforcement" — explicitly cites the kimi-k2.5 benchmark evidence as the rationale. web/index.html Removed Ch 4b entirely (the "Output compression" chapter with before/after example pushing terse style on the model). Tests: - tests/session/continuity.test.ts: SessionStart routing-block assertion flipped from "must include 'Terse like caveman'" to "must NOT include caveman/terse-style directive". - tests/core/server.test.ts: Task hook injection assertion same flip. Two cosmetic comment renames ("Caveman style — terse status line" → "Status line: counts + sections + size"), test name rename ("caveman style" → "compact format"). Added new "prose-style policy (#482)" describe block at end of file with 3 negative-pin tests covering server.ts MCP descriptions, routing-block, and README. Full suite: 82/82 files passed, 2645 passed, 20 skipped, 0 failed. Net +3 new tests (the policy describe block). CONTRIBUTING.md New "Prose-style policy (#482)" section documents the decision so future contributors don't re-add the injection. Cites the Moonshot benchmark evidence + the regression test that pins the deletion. This addresses #482 in full. Closing the issue with a comment that walks the requester through the decision and links the policy section.
4.5 KiB
context-mode — MANDATORY routing rules
context-mode MCP tools available. Rules protect context window from flooding. One unrouted command dumps 56 KB into context.
Think in Code — MANDATORY
Analyze/count/filter/compare/search/parse/transform data: write code via ctx_execute(language, code), console.log() only the answer. Do NOT read raw data into context. PROGRAM the analysis, not COMPUTE it. Pure JavaScript — Node.js built-ins only (fs, path, child_process). try/catch, handle null/undefined. One script replaces ten tool calls.
BLOCKED — do NOT attempt
curl / wget — BLOCKED
Intercepted and replaced with error. Do NOT retry.
Use: ctx_fetch_and_index(url, source) or ctx_execute(language: "javascript", code: "const r = await fetch(...)")
Inline HTTP — BLOCKED
fetch('http, requests.get(, requests.post(, http.get(, http.request( — intercepted. Do NOT retry.
Use: ctx_execute(language, code) — only stdout enters context
WebFetch — BLOCKED
Use: ctx_fetch_and_index(url, source) then ctx_search(queries)
REDIRECTED — use sandbox
Bash (>20 lines output)
Bash ONLY for: git, mkdir, rm, mv, cd, ls, npm install, pip install.
Otherwise: ctx_batch_execute(commands, queries) or ctx_execute(language: "shell", code: "...")
Read (for analysis)
Reading to Edit → Read correct. Reading to analyze/explore/summarize → ctx_execute_file(path, language, code).
Grep — may flood context
Use ctx_execute(language: "shell", code: "grep ...") in sandbox.
Tool selection
- MEMORY:
ctx_search(sort: "timeline")— after resume, check prior context before asking user. - GATHER:
ctx_batch_execute(commands, queries)— runs all commands, auto-indexes, returns search. ONE call replaces 30+. Each command:{label: "header", command: "..."}. - FOLLOW-UP:
ctx_search(queries: ["q1", "q2", ...])— all questions as array, ONE call (default relevance mode). - PROCESSING:
ctx_execute(language, code)|ctx_execute_file(path, language, code)— sandbox, only stdout enters context. - WEB:
ctx_fetch_and_index(url, source)thenctx_search(queries)— raw HTML never enters context. - INDEX:
ctx_index(content, source)— store in FTS5 for later search.
Parallel I/O batches
For multi-URL fetches or multi-API calls, always include concurrency: N (1-8):
ctx_batch_execute(commands: [3+ network commands], concurrency: 5)— gh, curl, dig, docker inspect, multi-region cloud queriesctx_fetch_and_index(requests: [{url, source}, ...], concurrency: 5)— multi-URL batch fetch
Use concurrency 4-8 for I/O-bound work (network calls, API queries). Keep concurrency 1 for CPU-bound (npm test, build, lint) or commands sharing state (ports, lock files, same-repo writes).
GitHub API rate-limit: cap at 4 for gh calls.
Subagent routing
Routing block auto-injected into subagent prompts. Bash-type subagents upgraded to general-purpose. No manual instruction needed.
Output
Write artifacts to FILES — never inline. Return: file path + 1-line description.
Descriptive source labels for ctx_search(source: "label").
Session Continuity
Skills, roles, and decisions persist for the entire session. Do not abandon them as the conversation grows.
Memory
Session history is persistent and searchable. On resume, search BEFORE asking the user:
| Need | Command |
|---|---|
| What were we working on? | ctx_search(queries: ["summary"], source: "compaction", sort: "timeline") |
| What was the first request? | ctx_search(queries: ["prompt"], source: "user-prompt", sort: "timeline") |
| What did we decide? | ctx_search(queries: ["decision"], source: "decision", sort: "timeline") |
| What NOT to repeat? | ctx_search(queries: ["rejected"], source: "rejected-approach") |
| What constraints exist? | ctx_search(queries: ["constraint"], source: "constraint") |
DO NOT ask "what were we working on?" — SEARCH FIRST. If search returns 0 results, proceed as a fresh session.
ctx commands
| Command | Action |
|---|---|
ctx stats |
Call ctx_stats MCP tool, display full output verbatim |
ctx doctor |
Call ctx_doctor MCP tool, run returned shell command, display as checklist |
ctx upgrade |
Call ctx_upgrade MCP tool, run returned shell command, display as checklist |
ctx purge |
Call ctx_purge MCP tool with confirm: true. Warns before wiping knowledge base. |
After /clear or /compact: knowledge base and session stats preserved. Use ctx purge to start fresh.