fix(context,tools): context hygiene + tool ergonomics from freestyle QA

Found while live-QAing freestyle_planning on a 12B local model:

- list_dir tool: recursive, .gitignore-aware listing so weak models stop
  flooding context with `ls -R` over node_modules/build/dist. Wired into
  the fileRead toggle + advertised to the planner (architect_freestyle).
- ContextClassifier: assistantToolCall turns are STRUCTURED, so the token
  pruner never shreds the model's own tool-call history — that was causing
  amnesia loops (re-issuing calls it had already made).
- Retire instruction-doc LLMLingua pruning (DOC_SOURCE_TYPES emptied): it
  fused load-bearing procedural text into unparseable soup. The static
  block stays small by dropping CLAUDE.md at the loader instead.
- AgentInstructionsLoader: load only AGENTS.md, not CLAUDE.md — the latter
  targets the outer assistant and polluted the agent's stage context.
- DefaultSessionReducer: WorkflowFailed flips session status to FAILED (was
  stuck ACTIVE forever, so clients/approval loops never saw a terminal).
- ShellTool: run shell command lines (cd/&&/pipes) via `sh -c`; unrunnable
  program is recoverable instead of an uncaught IOException killing the
  stage; malformed argv (non-string/collapsed-array) rejected with guidance.
- llmlingua sidecar: cap force_tokens to max_force_token (big docs blew the
  assert and 500'd, so doc pruning silently failed open).

Tests added/updated across all of the above.

examples/workflows/freestyle_planning.toml

@@ -10,7 +10,7 @@ start = "analyst"
 id = "analyst"
 prompt = "prompts/analyst_freestyle.md"
 produces = [{ name = "analysis", kind = "analysis" }]
-allowed_tools = ["file_read", "ShellTool", "task_search", "task_context", "task_create", "task_decompose"]
+allowed_tools = ["file_read", "list_dir", "shell", "task_search", "task_context", "task_create", "task_decompose"]
 token_budget = 16384
 max_retries = 2