Cline caps are hard-coded char budgets, not context-relative: read caps 48k chars / 2k lines / 2k chars-per-line documented exactly, one-line plan files marked never-read-back, README notes -c 131072 does not lift them

This commit is contained in:
Z User 2026-10-31 00:24:44 +00:00
parent 36c2dbf46e
commit 39a38c7c64
2 changed files with 22 additions and 9 deletions

View File

@ -299,12 +299,17 @@ output).
### Requirements
- **Local models through Cline (Windows 11 + llama.cpp):** Cline truncates file reads at
~50,000 characters, and cuts sooner when the model's context window is small — raise
your llama.cpp context (`llama-server -c 32768`) so a whole task fits. SKILL.md accounts
for this: the scripts are run, never read (`verify_plan_rules.py` is ~62 KB and exceeds
Cline's cap by design), and the big prompt files are read in bounded line ranges cited
in Steps 1, 3, and the short-stories section.
- **Local models through Cline (Windows 11 + llama.cpp):** Cline's file reads are capped by
hard-coded budgets that do not scale with the model's context window — about 48,000
characters and 2,000 lines per read, and any single line over 2,000 characters is cut
with a `[line truncated]` marker (command output and editor writes have similar caps).
Raising `llama-server -c` does not lift them: a 131072-token context runs this workflow
comfortably and still sees Cline's truncation notices. `-c` only needs to fit a whole
task (32768+ is plenty) and prevent the separate "request exceeds context" error.
SKILL.md works within the caps: the scripts are run, never read (`verify_plan_rules.py`
is ~62 KB and exceeds the read cap by design), the big prompt files are read in bounded
line ranges cited in Steps 1, 3, and the short-stories section, and one-line plan files
are copied or submitted with commands, never read back.
- `build_plan.py` and `verify_plan_rules.py`: Python 3 stdlib only — no venv packages
needed; the `.venv\Scripts\python.exe` form just matches a typical Windows project

View File

@ -30,9 +30,17 @@ never compose a scene from a guide the user has not approved.
- Start every task by running the smoke test under "Before you start". If it fails because
`h3-plan-gate` is missing files, stop and tell the user the folder is incomplete.
- **Never read `verify_plan_rules.py` or `build_plan.py`** — run them and work from their
printed output. Cline truncates file reads (~50,000 characters, sooner when the model's
context window is small), so the ~62 KB validator arrives cut no matter what. A
truncation notice is Cline's read cap, not repo corruption — do not "fix" truncated files.
printed output. Cline's read_file is capped by hard-coded budgets that do NOT grow with
the model's context window: about 48,000 characters per read, 2,000 lines per read, and
any single line longer than 2,000 characters is cut with a `[line truncated]` marker.
The ~62 KB validator arrives cut no matter how large `llama-server -c` is. A truncation
notice is Cline's read cap, not repo corruption — do not "fix" truncated files.
- **Never read back one-line JSON files** — `build_plan.py`'s output plan and
`h3_agent_workflow_examples/plan_oneline.json` are one-line JSON (the shipped example is
a single 5,807-character line), so a read always ends in `[line truncated]` at character
2,000 and the file looks broken. That is expected. Submit, copy, or move them with a
command; to inspect plan structure, read the pretty-printed `plan.json` in the same
examples folder instead.
- The two big prompt files are read in the **bounded line ranges** given in Step 1 and
Step 3 — never whole. If any read still arrives truncated, split it: read lines 1–230,
then 231 onward, and continue from where the last chunk ended.