…st) (#1323)

The system prompt was 24,387 bytes (≈ 6,100 tokens) — much of it
overlapping with the tool descriptions sitting right next to it in
the cache prefix. Sections like "When to propose a plan", "When to
ask the user to pick", and "When to track multi-step intent" each
recited rules that the tool's own description already carried.

Aggressive dedup pass:
- Drop the redundant "you have these filesystem tools" opening
  sentence — the API ships the tool list separately.
- Merge the three independent submit_plan / ask_choice / todo_write
  sections into one short "Picking the right tool" block.
- Fold "Exploration", "Trust what you already know", and "When the
  user wants to switch project" into shorter equivalents — same
  rules, no narrative.
- Collapse the foreground/background section. The full how-to lives
  in the run_command / run_background tool descriptions; the prompt
  only needs the picking rule.
- Compress the audit-mode rails (#610) prose around the six rails
  themselves. Every rail's load-bearing phrase is preserved verbatim
  so tests/code-prompt.test.ts still asserts on them.

Result: 24,387 → 11,956 bytes (-51%, ≈ 3,100 tokens per request).
Combined with PR #1320 / #1321 the cache-prefix tax per request is
now ~16k tokens instead of ~36k.

Behaviour unchanged — every rail / gate / mode constraint is still
asserted by the existing prompt tests.

Co-authored-by: reasonix <reasonix@deepseek.com>

…request) (esengine#1321)

Every byte of a tool's description and JSON-schema parameter blob ships
in every request. The spec list had grown to 39,377 bytes (≈ 10k tokens)
across 35 tools — much of it teaching guidance that was already covered
in the system prompt, plus verbose schema property descriptions that
restated what the parameter name already conveys.

Compress the heaviest twelve tools while preserving the actually
load-bearing rules (don't-use-for-A/B/C-menus on submit_plan, plan-mode
gate behavior on run_command, prefix-stable warning on remember):

  submit_plan          3517 → 1648  (-53%)
  revise_plan          2554 → 1482  (-42%)
  create_skill         2379 → 1367  (-43%)
  install_skill        2293 → 1510  (-34%)
  search_content       2167 → 1263  (-42%)
  mark_step_complete   1988 → 1037  (-48%)
  add_mcp_server       1898 → 1213  (-36%)
  run_command          1862 → 1376  (-26%)
  todo_write           1675 →  872  (-48%)
  ask_choice           1807 → 1313  (-27%)
  remember             1812 → 1306  (-28%)
  run_background       1611 → 1186  (-26%)

Total: 39,377 → 28,412 bytes (-28%, ≈ 2,740 tokens per request).

Behaviour unchanged — guidance that lived only in tool descriptions
(plan-mode interaction, ChoiceRequestedError stop semantics, prefix
re-load timing) is preserved verbatim. The system prompt and tool
error messages already carry the longer-form teaching content.

scripts/measure-tool-sizes.mts added so future audits don't have to
re-derive per-tool byte counts by hand.

Co-authored-by: reasonix <reasonix@deepseek.com>

…st) (esengine#1323)

The system prompt was 24,387 bytes (≈ 6,100 tokens) — much of it
overlapping with the tool descriptions sitting right next to it in
the cache prefix. Sections like "When to propose a plan", "When to
ask the user to pick", and "When to track multi-step intent" each
recited rules that the tool's own description already carried.

Aggressive dedup pass:
- Drop the redundant "you have these filesystem tools" opening
  sentence — the API ships the tool list separately.
- Merge the three independent submit_plan / ask_choice / todo_write
  sections into one short "Picking the right tool" block.
- Fold "Exploration", "Trust what you already know", and "When the
  user wants to switch project" into shorter equivalents — same
  rules, no narrative.
- Collapse the foreground/background section. The full how-to lives
  in the run_command / run_background tool descriptions; the prompt
  only needs the picking rule.
- Compress the audit-mode rails (esengine#610) prose around the six rails
  themselves. Every rail's load-bearing phrase is preserved verbatim
  so tests/code-prompt.test.ts still asserts on them.

Result: 24,387 → 11,956 bytes (-51%, ≈ 3,100 tokens per request).
Combined with PR esengine#1320 / esengine#1321 the cache-prefix tax per request is
now ~16k tokens instead of ~36k.

Behaviour unchanged — every rail / gate / mode constraint is still
asserted by the existing prompt tests.

Co-authored-by: reasonix <reasonix@deepseek.com>