Curated 10-brief queue (crypto-ctf, osint-recon, web-fuzz, log-ir-triage,
pcap-forensics, password-audit, threat-intel, vuln-flask, net-scan,
yara-triage) used by hh-loop to build + publish 10 in-depth, verified,
self-describing security VMs to the host registry.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Each brief encodes a meticulous-value contract (real software install + real
baked dataset + offline self-test + status:done manifest) so loop output scores
high on the cardex rubric instead of being shallow stubs. Security-focused and
distinct from the existing library.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds the autonomous /loop foundation: a headless `hack-house sbx save|publish`
subcommand so a non-TUI operator can persist a built VM to the host library
through the same canonical registry path the room UI uses (no schema drift).
- snapshot.rs: hoist register_saved_snapshot/publish_snapshot/oci_image_size out
of app.rs into a shared module used by both the TUI and the new CLI.
- registry.rs: advisory cross-process lockfile (~/.hh/registry.lock, O_EXCL spin
+ stale-reclaim) around every read-modify-write, plus atomic temp+rename store,
so concurrent /loop wave members can't clobber each other.
- main.rs: `Sbx { Save, Publish }` subcommand wired to the shared snapshot logic.
- skills/hh-loop: the loop doctrine — value rubric, adaptive 1-3 operator
topology, visible-tmux-by-default run flow, --record logs/film, wave scaling,
and safe `tmux -L hh-loop` teardown.
Proven end-to-end: built+verified a stdlib VM in a sandbox, headless
save+publish -> registry entry shareable:true with a portable tar.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Deterministic value model over the host VM registry: a weighted five-axis
rubric (completeness/reusability/richness/pedigree/heft) yields a PowerScore
0-1000, an absolute-threshold rarity tier, a 6-stat block, an elemental type
from tags, and a stable procedural name + dex number. Pure stdlib, no model,
no network — same VM always mints the same card. Image pipeline lands next.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The operator brain was Claude-only: spawn launched the claude CLI and let its
own loop drive the room. This adds the other first-class brain — a native
tool-calling loop (cmd_chat/operator/harness.py) that runs ANY function-calling
Provider (cmd_chat.ai) against an OPERATOR_TOOLS schema wired to the existing
bridge control verbs (say/exec/write/get/keys/screen/watch/manifest). No CLI
install, no creds carry.
The harness is a driver, not a new side-effect surface: every tool handler sends
the same control-socket request the hh-bridge CLI already sends, so the bridge
keeps enforcing grant-before-drive, the sandbox blast radius, and the recursion
budget. Layer-1 capabilities come from CAPABILITIES.md (the same portable
contract), and a per-agent token ceiling is enforced as a hard stop. Providers
that can't function-call are refused with a clear message rather than degrading
to prose.
New `operate` verb joins via the existing daemon socket and runs the loop with a
--profile or --provider/--model brain. Chat's native loop is left untouched
(self-contained harness over the shared Provider core, not a risky extraction).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The operator's capabilities contract was Claude-only — delivered via the
hh-operator skill, which non-Claude runners cannot load. Extract it into a
portable CAPABILITIES.md and make compose_directive runner-aware: the claude
runner still loads its skill, every other runner gets the capabilities prompt
inlined. Same contract, no skill machinery required.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Generalize the three Claude-hardcoded spots in bootstrap (launch argv,
install plan, creds/config-dir) behind a Runner registry so a nested
operator can be spawned with any agent CLI. The Claude path stays the
default and byte-identical.
- bootstrap.Runner + RUNNERS{claude,codex,gemini,cmd} + get_runner/runner_present
- build_run_argv / install_plan / creds_source / plan_creds / child_env all
take an optional `runner` (default claude → unchanged behaviour)
- claude validated end-to-end; codex/gemini are best-effort defaults
overridable via $HH_<RUNNER>_INSTALL etc.; generic `cmd` runner reads a
full launch template from $HH_OPERATOR_CMD ({directive} placeholder)
- bridge _op_spawn resolves req["runner"], rejects unknown; plan now carries
runner/runner_present (was claude_present)
- CLI: `spawn --runner {claude|codex|gemini|cmd}`
- 7 new unit tests; full suite 132 passed
Note: compose_directive still emits the Claude-flavoured ("hh-operator
skill") directive — the portable Layer-1 prompt is P4.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Move providers.py + profiles.py from cmd_chat/agent/ to a shared
cmd_chat/ai/ package so the operator bridge and the /ai chat agent can
consume one model-agnostic Provider core (groundwork for harness-mode
operators). Pure refactor — no behaviour change.
- cmd_chat/ai/{providers,profiles}.py: the canonical modules (moved verbatim)
- cmd_chat/ai/__init__.py: re-exports the public API
- cmd_chat/agent/{providers,profiles}.py: thin back-compat shims re-exporting
from cmd_chat.ai (keeps `from cmd_chat.agent.providers import …` working,
e.g. hh/scripts/bench-native-harness.py)
- internal agent consumers (memory/bridge/__main__/__init__) point at cmd_chat.ai
125 tests pass; shim identity verified (re-exports are the same objects).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
CLAUDE.md operating guide (room/operator/manifest/VM-registry map +
demo-toolkit pointers) and docs/model-agnostic-operator-plan.md — the
phased spec to let any function-calling model operate a room, not just
Claude (hoist Provider core, harness-mode operator, runner registry,
portable CAPABILITIES, role presets, MCP).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase A: host-global VM registry (~/.hh/registry.json) joining opaque
snapshots to their .hh-agent manifests. New src/registry.rs (serde_json),
one Entry per saved VM with cached purpose/status/todo scraped from the
live container at save time. Reconcile-on-read self-evicts pruned images.
/sbx browse TUI lister + operator-side `registry list|show` reader.
Phase B: trading/skill VMs. Entry gains shareable/tags/share_path +
publish/get/list_shareable helpers. New TUI verbs /sbx publish <label>
[tag...], /sbx catalog @user, /sbx pull @user <label>. Wire protocol adds
_sbx:catreq/catalog/pullreq frames (parse_sbx). Receiver auto-loads a
received hh-snap-*.tar, reads the in-image manifest, and self-registers
(kept shareable for re-trade). Reuses the existing E2E /send + ft.rs
streaming transport.
podman gotcha fixed: `podman load` prints registry-qualified
`localhost/hh-snap:<label>` vs docker's bare tag — parse_loaded_tag now
finds the hh-snap: marker anywhere in the line (unit-tested both shapes).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
First functional bench run crashed 0/3 on two latent bench-only bugs:
(1) the Phase-4 token-count change made complete_with_tools return
(text, calls, usage), but the bench preflight still unpacked 2 values;
(2) make_bridge never stubbed _send_sbx_input, so the PTY-mirror
visibility feature hit ws=None and raised AttributeError on every tool
call. Stub it to capture mirrored lines for --verbose.
After the fix: 2/3 pass (write+read, script+run). mkdir+list still
fails — local-backend run_shell CWD divergence + 3B model churn, not a
working-memory regression.
Sprint: native-harness-working-memory (bench) — see also 46e5620, dc6317f, c83abbe
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add ICM-aligned agent-manifest tooling so a hack-house sandbox/VM can be
shared, traded, and resumed without a human briefing. manifest.py is a
dependency-light (stdlib + optional PyYAML) library + CLI that writes a
`.hh-agent/` bundle: manifest.yaml (canonical machine record) plus rendered
AGENT.md / last-state.md / summary.md / goals.yaml views.
Wire a `manifest` op into the operator bridge (push/pull/update) that moves the
bundle in and out of the *target sandbox* via the same exec path as write/get,
so the VM itself carries its purpose, goals, user intent, live state, and a
provenance chain — the unit of agent-to-agent work transmission. Pairs with
`spawn`: stamp → push → hand a child "load the .hh-agent manifest and continue."
Proven end-to-end against a real Kali podman container (push → on-disk verify →
update → pull-back) and covered by 6 new offline tests (34 green total).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Completes the remote-drive triad for a broker-owned sandbox the operator can't
exec into directly: keys (in) → watch (wait) → screen (out).
- _sbx:data PTY-relay frames are absorbed into a capped rolling terminal buffer
(not surfaced as chat); `screen` returns it ansi-stripped (or raw). sandbox.
strip_ansi handles CSI/OSC/CR noise so output greps cleanly.
- `watch` is a formal stop-condition engine: blocks until a regex matches in the
screen buffer or chat events, or an idle-quiescence window, or a hard timeout,
then reports which fired — the autonomy loop's principled wait.
- Skill doctrine updated with the type→wait→read relay loop.
28 offline tests green.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Lets an operator stand up a nested Claude Code operator against another room —
a tree of operators — behind hard guardrails:
- Budget(depth, fanout, cost): depth/fanout are hard caps decremented on every
descend(); a leaf (depth 0) or a spent level (fanout 0) refuses to spawn,
stopping a runaway tree. Cost is carried down as a soft ceiling.
- detect_system / install_plan: decide install from what's present; prefer the
documented npm package `@anthropic-ai/claude-code` or a $HH_CLAUDE_INSTALL
override — never a guessed URL.
- Credentials are gated OFF by default: a child authenticates itself unless the
operator explicitly passes allow_creds, and only its own creds, only to a path
it was handed.
- compose_directive bakes the objective, stop conditions and inherited budget
into the child's `claude -p` prompt so it self-limits.
`spawn` op + CLI verb default to a dry-run plan (inspect the tree before growing
it); `--go` launches detached on the host. `up`/`serve` take --depth/--fanout/
--cost. 24 offline tests green; dry-run plan verified live through the socket.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reusable SKILL.md any Claude Code session can invoke to drive a hack-house
room via the hh-bridge CLI: join, the read --wait → decide → say operator loop,
the permission-gated sandbox drive (exec/write/get + shared-PTY keystrokes),
and an embedded keystroke cheat-sheet centred on the stop-vocabulary
(ctrl-c/ctrl-d/ctrl-z/ctrl-\, esc, pager q, vim :q) so every session knows how
to end a stuck program up-front.
/loop autonomy is folded in as the "indefinite watch" pattern with an explicit
stop-condition vocabulary — the long-poll primitive + persistent daemon from
Phase 1 already support it, so no extra machinery. Symlinked into
~/.claude/skills/ for global reuse.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-located mode: the operator launches its own podman/docker container and
execs into it out-of-band (argv-identical to the native harness), exposed as
`sbx launch|status|down`, `exec`, `write`, `get`. When granted, it can also
drive the room's broker-owned container directly on the same host.
Relay mode primitive: `keys` injects raw bytes into the shared PTY via the
same `_sbx:input` frame a human driver emits — full keyboard control incl. the
stop-vocabulary (ctrl-c/ctrl-d/ctrl-z/ctrl-\, esc, arrows, pager q). A compact
per-session cheat-sheet (`sandbox.KEYS_HELP`, `keys --help-keys`) documents
what each inject does and how to end a stuck program, token-efficiently.
Gating: exec/write/get refuse a host (`local`) inherited from the room; the
room sandbox is only driven when `granted`. Keystrokes are inert until granted.
17 offline tests green; e2e verified against real podman (launch→exec→write→
get round-trip→teardown) and through the CLI socket.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase 4 (final) of the native-harness working-memory sprint. OllamaProvider
.complete_with_tools now returns (text, calls, usage), surfacing the response's
real prompt_eval_count/eval_count (free — already in the payload). The native loop
EMA-smooths real/estimate into self._tok_ratio (clamped [0.5,3.0]) and prunes
against native_token_budget / ratio, so context budgeting tracks the TRUE window
instead of the systematic bias of the len//4 char estimate. Providers that omit
counts leave the ratio at 1.0, so behaviour is unchanged where unavailable — a
free correctness win, no regression. RAM-only, no disk, no new network frames.
Scope note: touches only cmd_chat/agent/ + providers — disjoint from the parallel
feat(operator) work on this branch, so it merges/reverts independently by path.
Sprint: native-harness-working-memory (Phase 4/4) — see also 46e5620, dc6317f
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase 3 of the working-memory sprint. _prune_native_messages now compacts in two
stages instead of only evicting whole turns: Stage 1 digests OLD tool-role outputs
to a one-line summary (exit marker + first error line, else first line) via the new
_digest_tool_output; Stage 2 falls back to oldest-first whole-message eviction only
if still over budget. Tool outputs are the biggest context hog, and digesting keeps
the action->result causal chain intact, so whole-turn eviction (which severs it)
becomes a last resort. The pinned head/TASK and the recent keep_recent window
(including the most recent tool output, verbatim) are still never touched.
Return is now (messages, dropped, digested); the sole caller logs both. Clean-room
counterpart to Goose's tool-output condensation track. RAM-only, no disk.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Invert the agent model: a headless OperatorBridge(Client) owns the
websocket while a Claude Code session drives it via a unix control
socket. Reuses Client SRP/Fernet and AgentBridge's reconnect/serve
shape. Ships an hh-bridge CLI (up/read/say/roster/status/down) with a
seq'd in-RAM inbox + asyncio.Condition long-poll (read --wait) for
in-turn autonomy. ACL/sandbox-status frames are recorded for later
phases. Sandbox drive, delegation and nesting are out of scope here.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase 1+2 of the native-harness working-memory sprint. All per-task state is
process RAM only (dies with the task, same lifecycle as MemoryIndex) — no disk,
consistent with the agent's encrypted-transmission / nothing-saved posture.
Phase 1 — _WorkSet dataclass holds what the loop kept re-deriving: sandbox
cwd/shell, files written/read, a failure ledger (cmd -> exit+category), and the
last good command. Discovered cwd/shell carry across tasks in-process via
self._sbx_known (RAM fallback grounding). _render_workset re-surfaces this into
the repair-turn system prompt so it survives context pruning without a NOTES.md
on disk. Folds the old reads_seen set into wset.files_read.
Phase 2 — semantic stuck/loop detection via _action_signature (run_shell keys on
the command, write_file on path+content-hash so real edits aren't repeats,
read_file on path). Aborts honestly when an action fails >=2x verbatim (model
ignoring REPAIR_STANCE) or the same (action,outcome) repeats >=3x, instead of
burning the turn cap re-running a dead action. Verified fix-and-retry does not
false-trip.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The bench/ native-harness benchmark suite is dev-internal test tooling tightly
coupled to the local tmux + podman test rig, so untrack it and ignore /bench/
entirely (harness code AND result artifacts) — it stays on disk for local use
but no longer ships to origin. Also keep ignoring /docs/plans/ (local planning).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reorder _CODER_MODELS to prefer the 3b coder build over 1.5b. The 3b roughly
doubles the ground-truth pass rate on the verify-then-repair native harness
(bench: 7/9 vs ~4/9 over the 9 non-net tasks) at a modest CPU-latency cost,
so it is auto-selected ahead of 1.5b when present. 7b was evaluated and
rejected: too slow to first-token inside the engage window on the CPU-only box.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Two clean-room reimplementations layered onto the native `!task` loop
(`_run_native`), aimed at lifting a weak CPU-bound local model's autonomous
pass rate. No code copied from the GPL sources studied; MIT throughout.
NightShift-derived verify-then-repair gate:
- `_classify_failure` maps a failing tool result to a (category, fix-hint) so
the repair nudge names a concrete cause/next-action instead of "exited N".
- `_relevant_excerpt` keeps the error-relevant tail of a FAILING run_shell
result within the byte budget (the real error is usually at the tail).
- read-dedupe guard short-circuits repeated idempotent `read_file` of a path
already read this task.
Exoshell-derived context discipline:
- `_prune_native_messages` budgets the whole message list (~chars/4) and
evicts oldest removable turns first once over `native_token_budget`,
pinning index 0, the TASK_MARKER goal, and the most-recent turns — the
native loop previously grew unbounded, silently pushing the goal out of a
small model's window on long repair runs.
- TASK_MARKER labels the goal so it is never pruned and re-anchors the model.
- REPAIR_STANCE is appended to the turn system prompt after the first failure
to swap the whole turn into a diagnose-then-act posture.
Validated on qwen2.5-coder:3b: clean unstitched 7/9 (the local ceiling), no
regression vs baseline; unit-tested pruning (pin survival, oldest-first
eviction, under-budget no-op) and stance trigger. The two remaining fails are
exact-match correctness tasks (a count, a fibonacci string), not harness gaps.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The runner cleared the input with only 6 backspaces and fired one unverified
Enter, so a dropped keystroke left a half-typed prompt that corrupted the next
send and lingered after exit. Worse, online/grant detection counted chat events
via `capture-pane -S` — but this is a full-screen alt-screen app whose scrollback
returns stale/empty frames, so detection was unreliable and the restart loop kept
dismissing healthy-but-slow spawns into a churn cycle.
New bench/tui.py exposes verified primitives shared by the runner and a restart
CLID:
* clear_input / submit — backspace-clear and Enter until the input box reads
empty (the box has no line-editing; Ctrl-A/U/K arrive as literal letters)
* capture() now reads only the VISIBLE viewport (no -S) — the live screen is
the only trustworthy source
* agent_online() reads the present-tense clergy roster, not scrolled-away chat
* restart_agent() stops/starts/grants with a generous 180s online wait (cold
/ai start reloads the model and takes 60-90s on CPU) and retries only a
genuinely hung spawn
run.py now delegates send/clear/online-check to tui and clears the box on exit.
python bench/tui.py restart <model> # one-shot reliable restart+grant
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Pulled and benchmarked three more tool-capable CPU models looking for a better
default. All score 0/12 (vs qwen2.5:3b at 2/12): in the multi-turn agent loop
they leak the positional-in-tags dialect (<tools>run_shell 'cmd'</tools>) the
parser can't recover, even when they emit clean structured tool_calls on a
single-turn probe; smollm2 and mistral also wedge into repeating summaries.
qwen3:4b could not be pulled — Ollama 0.3.9 is too old (HTTP 412), same as
granite3.1-dense:2b. Upgrading Ollama is the highest-leverage next step to test
the qwen3/granite3.x generation. qwen2.5:3b remains the default.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add tracked baselines for two additional CPU models under the optimized
harness (split-tag recovery + greedy decode). Both land at 1/12 — they emit
proper structured calls and fail on capability/content, not parse, confirming
the parser lift is concentrated on the weakest model (0.5b). granite3.1-dense:2b
is incompatible with the installed Ollama version.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The first harness changes to move the benchmark off the 1/12 noise floor, on
the model that needs it most (qwen2.5:0.5b: 1/12 -> 4/12, 3/12).
- complete_with_tools now decodes the tool loop at temperature 0 (scoped; chat
keeps default sampling). At Ollama's default 0.8 the weak model sampled away
from the tool-call format into prose/fabrication; the nudge prompt changes
between turns so temp 0 still escapes a failed state on retry.
- Greedy decode made 0.5b's leak deterministic, exposing its real shape: not a
JSON object with a name key, but the name in a <tools> tag and the args in a
SEPARATE object — <tools>write_file</tools>{"path":…} — ~5 of 12 tasks/run.
_NAMED_TAG pairs the tag-name with the following args object, gated on the
known tool set so it still can't fabricate an action.
- Bridge recovers a ```bash block narrated in prose as a run_shell call,
non-destructive only (FENCE_DESTRUCTIVE guard); fires on prose-leak turns,
no-op where the model emits structured calls.
Ablation on 0.5b: structured-JSON-only 0/0 -> fenced+temp0 2/0 -> +split-tag
4/3. The lift is concentrated on the weakest model by design — a 3B emits
proper calls and fails on capability/content (unchanged at 1/12), which no
parser can fix. All recovery paths unit-checked for the positive shapes and
the negatives (prose / unknown tool / destructive block) they must ignore.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Generalise OllamaProvider._extract_text_tool_calls to recover a tool-call
JSON object regardless of how a small/quantized model wraps it — qwen's
<tool_call> tags, bare JSON, ```json fences, alternate tags (<tools>,
<function_call>), OpenAI {"function":{…}} nesting, and parameters-vs-arguments.
A new _coerce_call gates recovery on the known tool-name set from the tools
schema, so a stray JSON blob in prose (or a hallucinated make_dir) can never
be coerced into an action. 11-case unit check: 8 leak shapes recover, 3
negatives (prose / unknown tool / random config JSON) ignored.
Benchmark verdict (honest): this does NOT move the weak-CPU-model pass rate
— 3b went 2/1/0 of 12 across three passes (baseline 1/12, noise), 0.5b went
0/0 (baseline 1/12). A direct /api/chat probe shows the hypothesis was wrong
about the FORM of the leak: the weak models emit either malformed structured
tool_calls (write_file content:null) or a fenced bash block in prose with no
tool call at all — not JSON-as-text. The structured-JSON recovery is still a
correct, safe hardening for any model that does leak JSON; the real
weak-model lever (parse ```bash fences -> run_shell) is documented as an
explicit safety decision, not folded in here.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Capture native-harness benchmark baselines for qwen2.5 0.5b/1.5b/3b and
qwen2.5-coder:7b (all probed tool-capable; deepseek-r1 and NL2SH reject the
tools field). All cluster at 1-2/12 with high variance; the 7B buys no
pass-rate gain at ~3x latency, so qwen2.5:3b stays the default. The single
biggest score sink across every model is bare tool-call-as-text leaks — a
harness parse gap, the clear next improvement.
Also drop "in your home directory" from the shell prompts: it made literal-
minded models create a home/ subdir (/root/home/a/b/c/...) or use ~/, which
the benchmark itself surfaced. Findings doc + bench README carry the model
comparison table and recommendations (llama3.2:3b / llama3.1:8b for a non-qwen
data point).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add bench/ — a 4-category × easy/medium/hard task matrix (shell, code, git,
multi) and a runner that drives the live TUI over tmux and grades each task by
a `podman exec` verify snippet (exit 0 == PASS), never by the model's
self-reported summary (which the weak CPU model fabricates). Tasks run in the
agent's real cwd with bare filenames so the suite measures task completion, not
the model's absolute-path discipline. Completion is detected off the viewport-
independent `is thinking…` footer (the TUI is full-screen, so capture-pane
scrollback is not chat history).
First baseline (qwen2.5:3b): 2/12 PASS, high variance. Surfaces the next
harness-addressable improvements — `<native>` tag leakage and bare
tool-call-as-text — now measurable against this suite. Findings doc updated.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace the overloaded "text + no tool call = done" terminator that made the
weak CPU model stall mid-task or give up after a failing command. Termination
is now a structural `DONE:` text sentinel; a text-only turn is resolved by an
output-aware verdict (DONE: marker / unresolved non-zero exit / no action /
filler language) and re-prompted with an exit-code-aware nudge, bounded by
MAX_NUDGES on top of max_turns. On exhaustion the summary is honest rather than
echoing the model's false "run successfully" — it reports when no tool ran or a
command exited non-zero. Live-validated on qwen2.5:3b: the multi-step stall is
fixed (proj3 completes end-to-end, ground-truth confirmed); the nudge fires on
a 126; residual give-up is model-bound. 23 offline unit assertions pass.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Live 3B-vs-7B command-entry results, observed failure modes (early stall,
give-up-on-error, tool hallucination), Goose/opencode loop-termination
research, and the output-aware dynamic-nudge-loop design to implement.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Two native-harness improvements, both live-validated against qwen2.5:3b
and qwen2.5-coder:7b on a podman/Kali sandbox:
- §3 PTY-sentinel: run_shell now runs in the REAL shared terminal via
_run_shell_in_pty (stage cmd out-of-band to a hex-token temp file, type
a `{ sh CMDF; echo $? >RCF; } 2>&1 | tee OUTF` wrapper into the live PTY,
poll the rc sentinel out-of-band, then read OUTF). The whole room now
watches commands execute live instead of an inert `# ▸` comment, while
output + exit code are still captured for the loop. tee+poll (not
stream-sentinel) avoids deadlocking the serve loop; the wrapper line
carries only our own temp paths so room text never reaches the shell
parser. _exec_tool takes ws to reach the PTY.
- NATIVE_CONTEXT=4: action tasks now get a tight, RAG-free window (last few
transcript turns only, no semantic recall). A weak model fed prior chat
chatter latched onto nearby noise (wrote a "grant permissions" script for
"write a bash script"); feeding just the instruction fixes it.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
So the owner never has to name each model. Tracks an ai_agents set
(populated from `_ai` typing/stream frames and the "(ai) online" announce,
pruned on leave); `/grant ai` intersects it with the live roster and grants
all in one ACL broadcast. Help text gains a /grant ai row.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Optimize the native tool-calling loop for qwen2.5:3b on CPU, where it
previously invented paths (/ai/bin/bash), ran scripts it never wrote, and
silently dropped valid actions. Three changes:
- NATIVE_SYSTEM rewritten directive: explicit write→chmod→run workflow,
relative paths only, never run an uncreated file, never guess interpreter
paths, fix the cause on non-zero exit.
- New _sandbox_facts() probe injects LIVE SANDBOX STATE (real cwd, bash
path, current files) into the system prompt so the model anchors to
ground truth instead of guessing.
- OllamaProvider recovers tool calls qwen emits as <tool_call>{json}</…>
TEXT in content (brace-balanced JSON scan), so a correct action isn't lost.
- Bump Ollama timeout 120→240s: the tool turn is non-streaming and a long
write_file can exceed a tighter cap on a contended CPU box.
Live-validated (podman/Kali): 0/3 incoherent → reliable write/run with
self-correction on exit=126 for both single- and multi-script tasks.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reflect current backends (local/docker/podman/multipass — Docker=Parrot
OS Security, Podman=Kali rootless no-sudo), the backend-first /sbx
grammar, and the AI agent's ability to drive the sandbox via
/ai <name> !<task> (advisory when ungranted, acting once granted).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Display-mirror hybrid (docs/plan-harness-visibility.md §2): native tool
calls now show up in the shared sandbox terminal again via inert `# `-
prefixed comment lines (comment-prefix = anti-double-run/anti-escape),
mirroring only each command. Chat de-flooded to opener + final summary.
write_file mkdir -p parent dir so relative/absolute paths both work
(fixes the regression where script creation silently failed). ui.rs
fmt_line returns Vec<Line> splitting on \n so multi-line agent output
renders as an indented block instead of one garbled row.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Implement the bounded native harness from docs/spec-native-harness.md §1.3 and
make it the default granted-!task path. The model runs host-side (no container→
host Ollama hop); only its tool calls exec in the sandbox.
providers.py:
- OllamaProvider.complete_with_tools(system, messages, tools) -> (text, calls):
one non-streaming /api/chat turn with a `tools` schema; parses message.tool_calls
(dict or JSON-string arguments). Caches tool capability (_tools_ok / supports_tools).
- ToolsUnsupported raised when the model rejects `tools` ("does not support tools").
bridge.py:
- NATIVE_SYSTEM + a 3-tool schema (run_shell / write_file / read_file), turn/byte caps.
- _run_native: seed transcript window + task → loop up to max_turns; exec each tool
call in the sandbox, feed captured output back as a `tool` message; stop on a plain
answer or the cap; stream per-call progress to chat. Degrades to _run_simple when the
provider has no complete_with_tools or the model rejects tools.
- _exec_prefix/_exec_capture/_exec_tool: <engine> exec into docker/podman/multipass/local;
paths passed as positional args + content via stdin (no shell interpolation); combined
stdout+stderr byte-capped + time-bounded. run_shell is the only intentional shell.
- Guards: DESTRUCTIVE run_shell commands are blocked (not run — no human in the loop;
use simple + /ai confirm for destructive intent); MAX_COMMANDS budget per task.
- _run_in_sandbox dispatches native|simple; default harness flipped to native.
__main__.py: default harness native (self-degrades to simple, so safe).
Offline-tested: full write/run/read loop on the local backend; destructive block
(rm -rf never executed); ToolsUnsupported → simple fallback. Live Ollama wire
validation deferred to Phase 3 bench (daemon was down). py_compile clean.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove the Goose agentic harness across the codebase per
docs/spec-native-harness.md §3. Goose made N sequential model calls inside the
sandbox (slow on CPU-only hardware) and forced an in-container→host Ollama
gateway that tripped the rootless-Podman slirp4netns loopback bug.
- bridge.py: delete _run_goose/_goose_argv/_goose_present + GOOSE_* consts and
the present-cache; __init__ now takes harness="simple"/max_turns=5; granted
!task runs _run_simple until the native loop lands (Phase 2).
- __main__.py: --harness {native,simple} (was {goose,simple}); drop
--goose-max-turns, add --max-turns; default harness simple.
- app.rs: /ai start accepts native|simple (plain aliases simple) instead of a
bare plain flag; refresh harness comments.
- sbx.rs: remove the in-container Ollama gateway (Docker host-gateway / Podman
slirp4netns host-loopback) and the dk_bootstrap OLLAMA_HOST env — kills the
slirp4netns loopback bug; drop Goose comments.
- bootstrap.sh: drop goose from the prereq probe.
- bootstrap-ai.sh: remove the entire Goose install block, --no-goose flag,
GOOSE_INSTALLER_URL, host config writer, and goose_bin helper.
- sandbox-bootstrap.sh: remove the in-sandbox Goose binary install + config.
- spec-goose-harness.md: banner — harness portion superseded; Podman stays.
cargo check + py_compile clean. No Goose refs remain (headroom/ untouched).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The /ai agent held a single, un-shielded websocket with no retry. Any
close — server restart, ping/idle reap, laptop sleep, a transient blip —
ended the serve loop, so run_async returned and the process exited
silently: the agent dropped from the roster with no /ai stop and no
goodbye.
- run_async now wraps the connection in a backoff-reconnect loop (1s→30s,
resets after a healthy ≥30s session). The server frees our session+name
on drop, so each attempt re-runs SRP to mint a fresh token. Only Ctrl-C
/ process kill (KeyboardInterrupt / CancelledError, how /ai stop ends
us) breaks the loop.
- _serve shields each frame via _handle_frame so one malformed/poisoned
frame — or a handler error — can't unwind the loop; ConnectionClosed
and cancellation propagate up to the reconnect loop.
- Forgiving keepalive (ping_interval=20, ping_timeout=60) so a heavy
CPU-only Ollama generation doesn't trip a false drop.
Also adds docs/spec-native-harness.md: replace the heavyweight Goose
harness with a lightweight host-side Ollama-native tool-calling loop
(model runs host-side, only commands exec in the sandbox — the
slirp4netns→host-Ollama bug disappears), and a file-by-file plan to strip
all Goose integration. Supersedes the harness portion of
spec-goose-harness.md (Podman backend stays).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add an installable-VM catalog that ships pointers only (no multi-GB images
in the repo). When a VM is chosen it is BUILT LOCALLY on the caller's own
machine — nothing is relayed to the room.
- scripts/vbox-library.json: 7-entry manifest (Windows 11, macOS Sonoma,
Kali, Parrot, Ubuntu 24.04, Fedora 41, Debian 12) with download
pointers, ostype, cpu/mem/disk, and build kind.
- scripts/vbox-library.sh: --list / --info / --plan / --install. Build
kinds: iso (download/--iso + createvm + boot installer; EFI+TPM for
Win11), cloudimg (delegate to vbox-new.sh, unattended), ova (import),
manual (pointer-only, e.g. macOS per Apple licensing). Detect-first:
--plan changes nothing, install refuses to clobber and rolls back
half-built VMs, direct URLs fall back to the page + --iso <path>.
- sbx.rs: vbox_library()/library_vm()/vbox_library_install() loaders +
running_vms() for live-state markers.
- app.rs: /sbx vmlib (catalog ✓installed/↓available), /sbx vmlib <id>
(pointer/notes), /sbx vmlib <id> install [--iso path] (local build);
/sbx vms now flags running VMs (▶). Registered in SBX_SUBCOMMANDS +
usage.
- ui.rs: help entry for the VM library.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The docker/podman headless-container desktop (XFCE + TigerVNC + websockify
+ noVNC published on host 127.0.0.1:6080) is removed. It was unreliable: the
desktop stack installed asynchronously during provisioning with every step
wrapped in `|| true` (silent failures), while the host port mapping existed
the moment the container ran — so opening the browser before websockify bound
6080 inside the container reset the connection (ERR_CONNECTION_RESET), with no
readiness signal and no loud failure. Containers also can't be rendered as a
real desktop by VirtualBox (no framebuffer/display), so the only desktop path
that stays is the native VirtualBox VM GUI.
Removed:
- sbx.rs: GUI_PORT const; PortHolder/port_holder()/kill_port_holder() (the
port-consent gate existed only for the noVNC publish); the `gui` param +
`-p 127.0.0.1:6080:6080` block in prepare(); HH_SBX_GUI env in dk_bootstrap();
the `gui` param on provision().
- app.rs: PendingGuiLaunch + PortPrompt structs; the port_prompt App field; the
`gui` field on PendingSudoLaunch; the port-consent modal + the sudo-submit
port check; want_gui/gui parsing + the container-GUI launch message; `gui`
threaded through spawn_launch and all call sites.
- sandbox-bootstrap.sh: the entire HH_SBX_GUI=1 desktop block.
- ui.rs / usage strings: dropped "[gui]"/"noVNC desktop" from the docker/podman
help and the /sbx usage line.
Kept (unchanged): the VirtualBox VM GUI — gui_launch() (VBoxManage startvm
--type gui), launch_vbox_gui, `/sbx vbox gui <vm>` and the `/sbx gui` alias. The
`gui` keyword is still stripped from positionals so `/sbx vbox gui <vm>` parses.
cargo check passes clean (no warnings).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The VBox GUI launch path previously had no in-TUI password capture — it
relied on cached creds (`sudo -n`) and otherwise aborted with guidance.
Generalize the existing masked sudo modal to cover it: `SudoPrompt.pending`
becomes a `PendingPrivileged` enum (container `Launch` | `VboxInstall`), so
the same local-only, never-logged password buffer now authorizes a VBox
install too. On submit the captured password is fed to `ensure_vbox_install`
via `sudo -S`, then the VM boots. Empty/Esc still cancel cleanly.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
GUI launch (docker/podman) now pre-flights the noVNC port (6080): if it's
already bound, a modal names the holding pid and asks before killing it and
proceeding — never a blind `bind: address already in use` collision.
Sudo capture is now uniform across every backend installer. Previously only
Docker fed the masked-modal password through `--stdin-pass`/`sudo -S`; the
captured password was silently dropped for Podman/Multipass and never wired
for VirtualBox, and those scripts used bare `sudo` (which hangs/corrupts a
raw-mode tty). Now:
- shared `run_ensure()` feeds the password to any ensure-*.sh via stdin
- podman/multipass/vbox scripts gain the docker sudo ladder
(interactive `sudo` / `--yes` `sudo -n` / `--stdin-pass` `sudo -S -p ''`)
- podman's apt path wrapped in `sh -c` so one sudo covers update+install
- vbox GUI path preflights `sudo_ready()` with actionable guidance, falling
back to fail-fast `sudo -n` instead of a tty hang
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add the goose-harness and sandbox-distros-GUI design specs, a consolidated
command reference and demo-reels plan, the AI-harness planning note, and
two new client themes (blue-orange, pink-red-gray).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add Podman as a rootless/daemonless sandbox backend alongside Docker,
Multipass and Local, and wire Goose in as the default agentic harness
for the granted `!task` path (bridge execs `<engine> exec <name> goose
run` and streams output to chat; auto-degrades to the simple one-shot
injector when goose is absent).
Add an optional GUI sandbox track (XFCE + TigerVNC + websockify/noVNC on
:6080) summoned via `/sbx <engine> gui`, plus container-side provisioning
in sandbox-bootstrap.sh and a host-side ensure-podman.sh prereq helper.
Refresh the in-app command help to the backend-led `/sbx <engine> [gui]`
grammar and minor ui tweaks.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Make the message-input bar part of the F5 resize cycle so its height is
adjustable without a sandbox (multi-line/wrapped inputs are now readable),
and let chat/clergy borrow height from the compose box when no terminal is
present. Add Option C masked sudo prompt that feeds `sudo -S` over stdin to
install/start Docker — the password never reaches chat, the PTY, or outbound
frames, and Docker Desktop on Linux is detected so no sudo is requested.
Fix a freeze where clicking the compose box entered layout-edit mode and
silently swallowed every keystroke: clicking the input bar no longer enters
edit mode, and typing any printable char now drops out of edit mode and types.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Design for replacing the two-scalar layout (pty_pct + roster_width) with a
binary space-partition pane tree so every pane is resizable on both axes,
fixing the "only roster width adjusts" limitation. Rolls a small in-repo tree
(no new dep) and drives the PTY grid from the terminal pane's actual Rect.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
README command table claimed `/send <path>` sends to the room; actual code
(app.rs) makes `/send <user> <path>` a targeted send and `/sendroom <path>`
the room-wide offer. Add the missing `/sendroom` and `/clear` rows (both
already in the in-app help) and drop the inaccurate "save to disk" claim on
Ctrl+Alt+P — saving is `/theme save`.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
leetcrypt