The entry rewrote every effort before the wire: `minimal -> low`, `high -> xhigh`,
`max -> xhigh`, under a comment claiming flash-server rejects anything outside
{xhigh, medium, low}. The engine says otherwise, in its own 400:
reasoning_effort 'maximum' unsupported; use minimal|low|medium|high|xhigh|max|none
and the harness passes an unmapped level through verbatim
(`compat.reasoningEffortMap?.[effort] ?? effort`, read in the pinned `dist/cli.js`). So the
fold protected nothing and silently changed what the caller asked for. The levels are not
cosmetic either — thinking length on one prompt at `max_tokens 2048`: `low` 766 reasoning
tokens, `xhigh` 614, `max` 1049 with the block closed at 1024.
The map is identity. Two things came out of capturing the wire (scratch HOME + a logging
proxy forwarding :7799 to :7770) that reading the config could not have told me:
- `:off` does not switch thinking off. The default disable mode sends the lowest effort,
so `off` left as `reasoning_effort: "minimal"` — and `minimal` still thinks, a canned
26-token block on every request. `reasoningDisableMode: none-effort` makes the harness
send `"none"`, which the engine honours as a real switch: `reasoning_tokens: 0`, answer
only, no thinking block rendered. Verified after the change, not before it.
- `:max` still arrives as `xhigh`: the harness clamps the requested effort to its own
ladder for this model before the map applies. Recorded in the comment as the engine-side
half of the contract rather than papered over — nothing in models.yml forces it.
Also pinned `top_k: 20` (accepted; the comment claimed the wire matched the server sampler
while `top_k` was never sent, leaving the recipe one env edit away from drifting), and
restored the pre-callsign model ids `frank-medium` / `frank-large` / `frank-small` — the
2026-10-08 rename had replaced the old id inside its own explanation, leaving "renamed
from frank-dragoon" three times and destroying the record needed to resolve an old id.
Verified: full suite green (0 failures), `kb-lint` 0 violations, YAML parses, and the four
efforts `off/high/medium/max` captured on the wire as `none/high/medium/xhigh` with
`top_k: 20` on every request. Refs #62
The one-work-log rule was written near-verbatim in four places — okf-spec,
okf-maintenance, log.md's preamble, and APPEND_SYSTEM.md. Four copies is four
authorities: each can be edited alone, and the next reader cannot tell which one the
linter implements. okf-spec.md owns it; maintenance now carries the action (collapse
a second log into the root, then re-run the linter) instead of re-arguing it; the log
page points at the rule rather than restating it. APPEND_SYSTEM keeps its imperative —
a system prompt cannot fetch a link at the moment of writing — but no longer carries
the duplicated reasoning.
`inventory/index.md` advertised "all userspace program configuration on the
workstation". The one page in that directory measures `~/.config` on blankpad: not
`~/.local/share`, not system package state, not the servers. A coverage claim that
names what was measured is the difference between a census and an alibi.
Refs #60
The live-sync heartbeat failed three hourly ticks (22:25, 23:27, 00:28) because a
managed skill that documents how to write a secret gate quotes the vendor's
all-zeros placeholder access-key id, and the shape rule fired on the
documentation. The machine tier is prose ABOUT credentials — the gate's own header
says a gate that fires there gets overridden until it means nothing — and a
heartbeat that cannot commit is a fleet that silently stops syncing config.
- shape hits are judged on the matched token, not the line: a documented fake is
inert, one live token sharing the file still blocks. The match is judged with
its following context, because a fixed quantifier truncates `EXAMPLE` to `EXA`
and the placeholder would vanish from what is examined.
- assignment values keep the stricter rule from the previous commit: a comment on
the same line cannot vouch for the token beside it.
- `ALLOW_SECRETS` remains the only override; nothing here widens what is scanned,
it only stops flagging text that was never a credential.
Verified against the payload that is stuck right now: the skill page that the
heartbeat cannot commit passes the fixed hook and was refused by the live one.
Refs #60
desk-1-home.kdl asked the Sceptre M34 for 3440x1440@239.999, a mode the panel
has never offered. niri does not reject an unavailable mode: it logs "configured
mode ... could not be found, falling back to preferred" and drives the EDID
preferred mode - 60 Hz - on every connector event, which is why the monitor
looked like it was dying rather than misconfigured. Read live with
`niri msg outputs`, the mode list tops out at 180.
Also destroys the copies of the phantom number: displays.md carried 239.999 in
its diagram, its panel table and its copy-paste pattern block, and its rule 4
claimed an impossible mode leaves the output dark. Rule 4 now states the real
failure mode, a new section separates config / DP-link / amdgpu-driver faults by
log signature, and wlr-randr is corrected to not-installed (three pages claimed
it; `pacman -Qs wlr` matches nothing).
Applied live with `niri msg output <panel> mode <mode>` - the temporary
per-output path, so no config reload and no logged-out session - verified at
3440x1440 @ 180.000 Hz with the compositor PID unchanged.
Refs #58
Rule 1 lost its work-log sentence in the point-of-use rewrite. Restoring it plainly would
re-create the old failure, where one agent kept logs in fifty-odd files and none of them
was the place to look. So the restored rule names the file: a single page in the
knowledge base — omp-docs/log.md here, whichever change-log page a repo's index names
elsewhere — read before starting, written in the same commit as the change it describes,
under that page's own entry law. Explicitly banned: per-directory logs, per-project or
per-run logs, NOTES.md/PROGRESS.md/STATUS.md/worklog/, dated diaries. If a KB has no log
page, the fix is one at its root plus an index entry, never a log somewhere else.
The bundle's own log records the change and the two-log incident it came out of.
The RDNA4 offload campaign left three HIP probes and a driver sitting loose in
$HOME, where the next agent cannot find them and a stray `rm -rf` ends them. They
answer a recurring question on the R9700 box — is decode limited by the link or by
the engine's read path — so they move into scripts/probes/ with a driver, tests,
and a page.
Two things changed in the code, not just the address:
- The driver measures only after a guard that runs BEFORE anything is built:
gpu_busy_percent over PCIE_PROBE_MAX_BUSY, or any process holding /dev/kfd, is a
refusal (exit 3) rather than a measurement. A benchmark taken beside a live
endpoint is both disruptive and worthless, and silently unloading a production
service is not a tool's decision — --force is the human saying otherwise. The
refusal-before-build ordering is pinned by a test, because the sysfs and kfd
paths are overridable and a refusal is otherwise untestable off the card box.
- hip_smoke.hip and pcie_probe.hip called hipGetDeviceProperties, which SEGFAULTS
on this ROCm build even though hipGetDeviceCount works; only pcie_bw.hip avoided
it, and that is the one that produced the campaign's numbers. Device identity now
comes from the sysfs topology block the driver prints, so all three probes are
runnable on the build we ship. Verified here: refusal on a real busy APU (exit 3),
missing toolchain (exit 4), missing sources (exit 2). Compile-and-run still needs
the box with hipcc.
Refs #53
Three things were in history that should not have been, all invisible to the
tool that deploys them:
- `home/dank/.omp/dotfiles/**` — 3 files re-hosting `niri/cfg/autostart.kdl`,
`niri/cfg/keybinds.kdl` and `systemd-user/…/swayidle-lock.service` under a
prefix `scripts/deploy.sh` never reads (it walks `dotfiles/<pkg>/`). The live
package is deployed and symlinked, so the copy asserted nothing; it only
misled. Moved to `.AGENTTRASH/20261020-tracked-debt/`.
- `.pre-pull-backup/**` — 2 managed-skill files, a local backup in history,
which `.gitignore` itself declares must stay on disk only. Same destination.
- `dotfiles/noctalia/.local/share/noctalia/wallpapers/` — 23 images, 80.4 MiB.
Untracked, files left in place: `~/.local/share/noctalia/wallpapers` symlinks
into this directory and the theme reads it, so deleting them would have gone
black. Every filename carries its wallhaven ID, so a fresh box restores the
set with one `curl` per file plus `deploy.sh`; losing the set is an acceptable
outcome, 80 MiB per clone forever is not.
New page `omp-docs/tools/dotfiles-deploy.md` owns the deploy contract (the
`--check`/sync/`--adopt` modes, `zz-$(hostname)` override semantics, the dconf
and systemd-user special cases) because nothing documented it, plus two laws:
media the pipeline consumes is content not config, and `systemctl --user
enable` snapshots an absolute path — so enabling from a worktree means the
seven-day reap silently kills the unit while `is-enabled` keeps answering
`enabled`. That second law is not theoretical:
`default.target.wants/blankpad-desk-autoswitch.service` pointed into the reaped
`wt/de-redesign-20261009` and desk auto-switching had been off since. It is
relinked to `dotfiles/zz-blankpad/` and left inactive rather than started mid-
session; starting it applies a desk profile.
Refs #40
The Matt Pocock engineering skills (triage, to-tickets, to-spec, research,
implement) assume per-repo config for three things: where work items live, what
the triage label vocabulary is, and where domain docs sit. Nothing here declared
any of them, so every skill run re-guessed.
Frank's decisions, 2026-10-20: Gitea on Dankclaw as the tracker (no gh, no tea on
any box, so it is recorded as prose — auth via a 600 curl config, the endpoint
table, the 404-on-empty-token / form-encoded-merge / empty-200-on-merge traps,
and the AGENTS.md laws a skill must not renegotiate); the canonical triage label
vocabulary kept as-is, with the five labels created on the tracker (the repo had
zero, so nothing collides); single-context domain docs.
The configuration moved out of the skill's default docs/agents/ and into the one
knowledge base this repo allows: three pages under omp-docs/skills/, an empty
omp-docs/decisions/ carrying the ADR format rule, an "## Agent skills" pointer
block in AGENTS.md, index entries, and a log entry. The glossary is deliberately
not created — every candidate term already has an owner in system/, core/ or
skills/, and a term list that paraphrases them is a second owner of facts already
written down. Known cost: re-running setup-matt-pocock-skills will not recognise
omp-docs/skills/* and will offer to recreate docs/agents/; the AGENTS.md pointer
is what stops that.
Also closes a verification gap found on the way: kb-lint.bats only ever points
the linter at fixtures, so a page with no frontmatter or an index that stops
listing a file passes the suite silently. Two tests now run the linter and the
coverage checker against the shipped omp-docs/ tree.
Refs #18
The Limits bullet still described the pre-fix behaviour ("thinking and the
final message only"), which dank/dankvoice#3 just retired: narration rides
`turn_end`, so the assistant's prose and thinking are spoken while the agent
works. Records the two rules an operator needs when audio seems missing —
arming is read at startup only, and a turn landing during playback is dropped
rather than queued — plus where the design rationale lives.
kb-lint 0 violations, index-coverage 0 issues.
The same bytes are already in origin/main via PR #37; the workspace policy here
blocks restore/stash/switch, so the identical working-tree copies are committed
rather than discarded, letting the deployment take the merge commit.
The extension's full behaviour now lives in dank/dankvoice (41/41
black-box suite against the built artifact; this branch's 26 drop-in
tests are a strict subset, names ported 1:1). Delete the drop-in and
its bats suite; repoint omp-docs/core/dankvoice.md at the package,
keeping only fleet wiring (frank-tts + @tiny resolution, the
enable_thinking hybrid law, voice/roster contract) and the cutover
procedure. kb-lint 0 violations; bats tests/ 45/45.
Refs #34
Refs #32
The humanizer's raw fetch to frank-zealot omitted
chat_template_kwargs.enable_thinking: false, so the Qwen35 hybrid thought
instead of answering (6.9 s reasoning_content, empty content - timing out
the 6 s abort; or answering "" and muting the unit). Pin the kwarg, treat
empty-content-with-reasoning as failure over silence, and stop the menu's
test phrase from blocking its handler. Three red-captured regressions;
26/26 green, verified red against HEAD.
Changelog entries were essay-length incident writeups duplicating the
concept pages they created. New entry law recorded at top of log.md:
one dated entry = what changed + durable lesson + concept-page link;
full narratives recoverable via git log --follow. Every date, headline,
incident, and durable lesson preserved; narrative prose that already
lives on concept pages (caddy handle-ordering, gluetun/limiter field
notes, quadlet gotchas, faillock law, credential-store trap) replaced
by pointers. Historical pi-era entries tagged '(pi era)'.
Also: self-hosted-web-agents.md '44.5% peer-reviewed CUA' figure could
not be re-attributed to any published benchmark via web search today
— dropped with an explicit note; the valid core claim (87.4% = managed
Cloud, not the self-hostable library) kept.
Refs #29
Two directories documented one concept (pi npm packages): top-level
packages/ (browser-use, permission-system) and pi/packages/ (peon,
web-access, subagents). Merged into pi/packages/ — one row per package
in a single index, detail pages kept as archive records. The root
index's 'OMP packages and skills' section was doubly wrong (packages/
was pi, not OMP); it becomes a Skills tooling section pointing at the
consolidated /skills/ tree.
Moved the two LIVE omp pages out of the retired-harness banner:
pi/statusline.md -> core/statusline.md, pi/subagent-limits.md ->
core/subagent-limits.md (both document omp behavior against
agent/config.yml + the installed bundle; pi/index.md itself admitted
the mis-shelf). Retargeted every inbound link (core/index, system/
dankserver, log x2, pi/index, settings x2, moved files' own xrefs).
Refs #29
Answers spoken with /say were verbatim markdown-cleaned text while the
built-in vocalizer speaks prose rewritten by a small model. Port that
enhancer (bundle Qyr/T6o: same prompt, 1536 tokens, 6 s, 4000-char
window, null-fallback/empty-skip contract) and resolve its model from
modelRoles.tiny exactly as the harness does, so auto-speak sounds like
the enhanced TUI voice instead of a pager. agent_end now also speaks the
last assistant message's thinking blocks as their own unit before the
answer; narration-flagged thinking counts as answer text. Both knobs are
menu-toggleable, persisted, and env-overridable; the fixture gains a
chat route (echo/prose) and seven scenarios cover the round-trip, the
404 fallback, ordering, the kill switches and role resolution (22 green,
58 suite-wide; live smoke through real frank-zealot + frank-tts).
Refs #27
agent_end awaited one whole-answer synthesis; long answers hit the
harness 30s extension-handler timeout (handler timed out after 30000ms)
and the single frank-tts request was the stall. Port the built-in
speak-aloud vocalizer's utterance segmenter (fHe in dist/cli.js) so
segmentation matches the known-working stock path: [.!?…]-run
boundaries, abbreviation/decimal/inline-code guards, fenced code and
tables swallowed, markup cleaned. Play segment N to player exit before
synthesizing N+1 (audio never stacks); fire-and-forget the chain so
utterances outlive the handler timeout; a run token retires superseded
chains. DANKVOICE_MAX becomes a per-answer budget of whole segments;
stripForSpeech is gone. Refs #23 follow-up.
Bare /say and /say menu open a ctx.ui.select settings surface: auto-speak
toggle, voice from the frank-tts roster, PipeWire output device
(pw-play --target / paplay --device=), player binary, spoken cap, and a
test phrase. Choices persist to ~/.omp/cache/dankvoice.json (env wins
over file); detectPlayer becomes availablePlayers + playerFor. Fixes
the silent-route failure mode where synthesized audio played into an
offline Bluetooth sink. 15/15 dankvoice bats, 51/51 suite, live pty
TUI smoke armed the footer and wrote the settings file.
Every speak-aloud path in the harness gates `--voice` against the bundled
Kokoro enum, so the frank-tts cloned voices could only ever produce files.
`agent/extensions/` auto-loads, so one tracked extension adds the missing
transport: `/say <text>`, `/say on|off` (narrate the final answer at
`agent_end`, silent when the harness schedules a continuation),
`/say voice|voices|stop|status` — OpenAI speech POST plus playback through
pw-play/paplay/ffplay. The endpoint and served id come from the
`api: openai-speech` row of `agent/models.yml`, so the file carries no host
and no secret; markdown is stripped, over-long utterances are cut at a
sentence bound, and the temp WAV dir is always removed.
Verified live in a real TUI session on the workstation (roster
narrator1/narrator2/narrator3; 2.08 s and 1.44 s WAVs played; auto-speak
spoke each finished answer, 1.20 s WAV; unknown voice answers with the
roster). Black-box seam `tests/dankvoice.bats` (12 tests) drives the
extension against a fixture speech endpoint and a stub player.
cwd: /home/blank/.omp/wt/feat-21-dankvoice
session: wt-20261017-181502-6d1fc1f
Documents the 16:46 layout loss (uncommitted GUI state stashed, stash
entry destroyed during trunk sync) and the git-fsck recovery landed in
c87b378. Lesson codified in managed skills.
Every agent-opened issue/PR body ends with cwd + omp session id so machine
filed artifacts trace back to the box and session that produced them - the
live-sync daemon and filing agents make 'who/what wrote this' a real
question now. Mirrored in dank/serverAppConfigs by its own PR.
Refs #11
The checkout IS the deployment and symlinks make live state (memories,
auto-learned skills, agent config, desktop settings) dirty tracked files
continuously; no commit law covered machine churn, so it accumulated
unpushed for weeks — one disk held the only copy of ~61 managed skills,
a settings fix stayed local-only, and dirty trees refused merges.
scripts/live-sync.sh (systemd user timer, ~15 min) commits ONLY the
allowlisted machine/desktop-live tiers with hooks armed, merges
origin/main, pushes, and reaps wt/* worktrees older than 7 days. Task
dirt outside the allowlist is never staged; conflicts abort and alert
via ntfy; branch guard refuses off main; never --no-verify, never force.
Black-box contract pinned by tests/live-sync.bats (9 tests, suite 36/36).
Refs #11
Trunk-based SDLC gains the delivery habit: xdg-open the PR URL the moment
it's filed (mirrors dank/serverAppConfigs step 6, riding their PR #6).
Changelog entry added.
Single-dev repo: develop bought nothing and cost a back-merge every
release. New law: branch off freshest origin/main per work item, PRs
target main directly, --no-ff merges kept, tags vX.Y.Z straight on main.
No release branches, no back-merges. Changelog entry added.
Census against serverAppConfigs-docs + read-only ssh probes on both boxes
(2026-10-18). Corrections to OURS, flags only for THEIRS:
- Hardware drift: RTX 3090 moved dankserver->dankclaw 2026-09; dankserver
now 2x Radeon AI PRO R9700 (Ryzen 7 7700X stays), 246 GiB RAM, no swap;
dankclaw 121 GiB RAM + 64 GiB swap. Served: frank-archon radiance TP=2
(:7770 dankserver), frank-dragoon halogen NATIVE v2 (:7770 dankclaw
iGPU), frank-zealot llama.cpp PrismML Ternary-Bonsai-2-27B PQ2_0
(:7771 dankclaw, 3090). Kernel 7.0.0-38 both; dankclaw :7770 no auth.
- Gitea: gitea-caddy quadlet retired (fleet Caddy fronts the domain);
data tree moved to /home/dank/Containers/gitea/ (2026-09-01 standard),
/opt/gitea is a do-not-use leftover; LAN IP + live domain.
- pi-era pages (core/settings, pi/models + role tables) demarcated as
ARCHIVE with banners; llama-cpp fleet note modernized.
- AGENTS.md machine-facts table corrected (it still routed prod warnings
at frank-default-large).
Their-side staleness flagged in the changelog entry for the operator,
not edited: OMP Harness thresholdPercent 70 (live is 95), Ubuntu 24.04
OS table, 2026-09-03 stack mapping, superseded pi-KB backlink.
Refs #5
Full-source re-verification of pi/statusline.md against dist/cli.js plus a
live screenshot matrix. False and corrected: claude's top rule is NOT the
context meter (meter is box/band only; rule elsewhere is a session-accent
filler); 'Tern-only' composer-bar claim (real knob: composer.tokenRate);
segmentOptions gated to preset custom (they merge under every preset - only
the arrays are gated); mode segment yolo (states are Plan/Prewalk/Goal/
Vibe/Loop). Expanded: 7-preset defaults table, per-shape statusAttachment/
bottomBar model incl. plugin shape registry, contextLine per-value behavior,
27-segment catalog with triggers and options, segmentOptions schema, full
statusLine.*/composer.* knob registry. Changelog entry added.
Refs #5
Left group keeps the where-am-I identity (host, path, git, pr); the right
cluster carries what's-running (model) plus cost and elapsed time, painted
in array order. context_pct verified to hoist into the inline meter spacer
regardless of array position. Docs + changelog updated.
Band merges the segment groups into one row above the editor and paints
the labeled inline context meter (60% - 200K); the session title rides
the working row. Cumulative token totals live in /session - off the bar.
Docs and changelog updated to the band final state.
Named presets silently ignore leftSegments/rightSegments - the nerd
preset's baked arrays were what we saw, and the configured arrays were
dead. preset: custom with explicit arrays: chip (right) = token_in,
token_out, cost, context_pct, time_spent; strip (left) = session_name,
vim, mode, hostname, model, path, git (branch text off - worktree
branches are autogenerated; dirty badges kept), pr. Cut tok/s duplicate,
clock, cache trio, context_total, subagents, session hash. Composer
shape stays claude after live box/band/rule trials. models.yml display
name frank-archon drops the (dankserver) suffix; id untouched.
Docs: pi/statusline.md records the preset-vs-custom trap, shape->segment
routing, and the context-meter truth.
Five layers: live-value match across all 77 revs / 765 reachable blobs; the whole
object store including the 2 dangling blobs (one synthetic sk-shaped fixture,
never committed on any ref); 21 provider shapes with every hit triaged; an
entropy scan of HEAD; all 23 git-lfs payloads (81 MiB) scanned as bytes.
Gate gap noted: hf_/npm_/pypi-/AIza/basic-auth-URL shapes are absent from
.githooks/pre-commit. Quoting the fixture literal in this commit message was
rejected by the gate, so it is described instead.
Also chmod 600 ~/.config/secrets/hf_token_write (was world-readable).
New tools/gitea/client-access.md: the credential file (~/.config/secrets/gitea,
0600) behind a global 'store --file' helper, the git 2.56 credential-store trap
(the key=value block shape parses to an empty get; the file must be one
http://user:token@host:port URL per line), the GIT_TERMINAL_PROMPT=0
verification loop, the lfs.locksverify key that silences the URL echo, and the
rotation procedure. Cross-linked from the gitea index and management pages.
Both workstation remotes are now bare URLs; the token lives in one file.
Callsign renames verified against the live endpoints (all three accept the
new request ids and return reasoning_content), compaction endpoint moved and
22 model references checked for integrity, runtime scratch quarantined in
.gitignore, skill libraries shown to be renames with history intact at
12c5e88.
Second half of the purge (the first covered the persona, AGENTS.md, README.md,
and the distillation of the per-host pages). Every remaining identifier is
gone: hostname, both LAN addresses, and the Wi-Fi SSID.
34 count-checked substitutions across 22 files — indexes, tool/research pages,
the inventory census, the changelog, one generated HTML report, and five
managed skills. Facts kept, names dropped: a history entry keeps its four-page
listing as a nameless summary, the syslog excerpts keep every timestamp and
line, and the census keeps the .zcompdump rows (the zsh version stamps are the
useful part, the hostname was not).
Deliberately unchanged: install.sh's --host case list and the
packages/<hostname>.pkgs filename, where the hostname is the lookup key.
Left out of scope by Frank's call: agent/memories/** — past-session records the
learn pipeline rewrites.
kb-lint: 0 violations. Residual grep for the hostname, its two IPs, and the
SSID across omp-docs/ and agent/managed-skills/: no hits.
Copy of the legacy ~/.pi/agent/docs tree rebuilt as the command-center
knowledge base: concept files promoted to path identity, index listings
stripped of frontmatter, all wikilinks/relative links resolved, every
checkable claim verified against the live fleet (Dankserver = Ryzen 7
7700X + RTX 3090; Strix Halo is Dankclaw), unverifiable claims tagged.
kb-lint exits 0.
dank
Frank