Skip to content

akiiwa-games/AgentCompass / skills/agent-compass

agent-compass

Unverified

Multi-agent task router and subscription quota scheduler. Choose owner, model and effort, outbound authorization, advisor use, and continuation for delegated work. Reads local usage and effective account policy, profiles the task with the configured System-One scorer, then allocates a compatible worker. Use for independent worker or advisor work; ordinary local execution needs no routing.

FileKindSize
SKILL.mdskill md5.6 KB
agents/openai.yamlconfig427 B
calibration/README.mdmarkdown55.2 KB
calibration/arena_capture.pyscript14.0 KB
calibration/community.pyscript25.2 KB
calibration/five_axis.pyscript196.4 KB
calibration/measurements.pyscript68.4 KB
calibration/refresh.pyscript48.6 KB
calibration/series.pyscript10.2 KB
calibration/sources.pyscript24.1 KB
calibration/tiers.pyscript14.6 KB
grok-bot/.gitignoreother47 B
grok-bot/dependency_patch.pyscript4.5 KB
grok-bot/install_cli.pyscript3.6 KB
grok-bot/test-recovery.mjsscript8.6 KB
grok-bot/test_install_cli.pyscript4.1 KB
grok-bot/upstream.jsonconfig3.4 KB
references/accounts.jsonconfig155.9 KB
references/classifier-cases.jsonconfig21.5 KB
references/dispatch-lifecycle.mdmarkdown37.7 KB
references/dispatch-rules.mdmarkdown11.3 KB
references/effort-matching.mdmarkdown5.1 KB
references/fast-modes.mdmarkdown3.5 KB
references/five-axis-scores.jsonconfig3.1 MB
references/gemini-decisions-api.mdmarkdown2.9 KB
references/grok-bot.mdmarkdown13.0 KB
references/herdr.mdmarkdown5.1 KB
references/jev-api.mdmarkdown1.6 KB
references/level-3-conversion.jsonconfig2.4 KB
references/model-capabilities.jsonconfig3.6 KB
references/model-capabilities.mdmarkdown14.6 KB
references/parent-address-gate.mdmarkdown10.1 KB
references/plan-templates.jsonconfig13.2 KB
references/provider-discovery.mdmarkdown11.5 KB
references/provider-usage.mdmarkdown4.1 KB
references/public-data.jsonconfig816 B
references/public-measurements.jsonconfig1.3 MB
references/roll-reference.mdmarkdown26.7 KB
references/router-internals.mdmarkdown94.5 KB
references/task-classifier.mdmarkdown6.1 KB
references/user-scores.jsonconfig6.8 KB
references/work-kinds.mdmarkdown4.7 KB
references/worker-communication.mdmarkdown2.1 KB
scripts/account_fleet.pyscript32.9 KB
scripts/account_state.pyscript22.2 KB
scripts/accounts_override.pyscript14.8 KB
scripts/allocator.pyscript19.8 KB
scripts/antigravity_plan.pyscript18.7 KB
scripts/antigravity_telemetry.pyscript5.1 KB
scripts/auto_bundles.pyscript31.3 KB
scripts/axis_scores.pyscript40.3 KB
scripts/caller_input.pyscript3.0 KB
scripts/candidate_eval.pyscript54.0 KB
scripts/candidate_ranking.pyscript9.5 KB
scripts/capability_catalog.pyscript8.0 KB
scripts/check_router_contract.pyscript2.8 KB
scripts/conftest.pyscript2.5 KB
scripts/cwd_policy.pyscript1.5 KB
scripts/data_age.pyscript1.5 KB
scripts/decision_ledger.pyscript18.9 KB
scripts/demo_plan.pyscript6.3 KB
scripts/dispatch_fallback.pyscript40.3 KB
scripts/dispatch_flow.pyscript63.1 KB
scripts/dispatch_ground_truth.pyscript43.5 KB
scripts/dispatch_registry.pyscript130.8 KB
scripts/entry.pyscript1.7 KB
scripts/env_vars.pyscript580 B
scripts/fast_modes.pyscript2.5 KB
scripts/fixtures/grok-api-responses.jsonconfig907 B
scripts/fixtures/reference-accounts.jsonconfig26.9 KB
scripts/grok_bot_paths.pyscript1005 B
scripts/host_herdr.pyscript16.8 KB
scripts/host_placement.pyscript9.4 KB
scripts/host_policy.pyscript10.1 KB
scripts/launch_failures.pyscript5.9 KB
scripts/launch_values.pyscript12.3 KB
scripts/ledger_snapshot.pyscript11.1 KB
scripts/live_models.pyscript21.1 KB
scripts/model_latch.pyscript7.6 KB
scripts/model_resolution.pyscript18.0 KB
scripts/override_store.pyscript2.2 KB
scripts/plan_state_residue.pyscript5.0 KB
scripts/provider_inventory.pyscript128.2 KB
scripts/provider_retry.pyscript34.6 KB
scripts/public_data.pyscript59.2 KB
scripts/quota_economy.pyscript38.0 KB
scripts/quota_exhaustion.pyscript5.9 KB
scripts/quota_overrides.pyscript25.6 KB
scripts/quota_pools.pyscript29.8 KB
scripts/reference_provenance.pyscript32.3 KB
scripts/released_names.pyscript3.8 KB
scripts/roll.pyscript106.3 KB
scripts/route_judge.pyscript83.5 KB
scripts/route_simulation.pyscript26.3 KB
scripts/router_test_support.pyscript6.7 KB
scripts/run_grok_bot.pyscript18.8 KB
scripts/scorer_client.pyscript55.2 KB
scripts/stage_help.pyscript5.8 KB
scripts/state_upgrade.pyscript24.0 KB
scripts/subscription_cli.pyscript3.6 KB
scripts/task_intake.pyscript33.6 KB
scripts/test_account_identity.pyscript7.1 KB
scripts/test_account_state.pyscript63.3 KB
scripts/test_accounts_override.pyscript30.1 KB
scripts/test_allocator.pyscript94.2 KB
scripts/test_antigravity_plan.pyscript13.2 KB
scripts/test_applicable_quota.pyscript9.1 KB
scripts/test_auto_bundles.pyscript10.8 KB
scripts/test_billing_routing_policy.pyscript9.9 KB
scripts/test_cadence_count.pyscript32.2 KB
scripts/test_calibration_units.pyscript27.0 KB
scripts/test_capability_catalog.pyscript5.0 KB
scripts/test_decision_ledger.pyscript8.5 KB
scripts/test_demo_plan.pyscript3.4 KB
scripts/test_derived_state.pyscript21.6 KB
scripts/test_dispatch.pyscript352.4 KB
scripts/test_dispatch_answers.pyscript14.8 KB
scripts/test_dispatch_fallback.pyscript53.2 KB
scripts/test_entry.pyscript1.6 KB
scripts/test_fast_modes.pyscript9.9 KB
scripts/test_five_axis.pyscript290.1 KB
scripts/test_five_axis_economy.pyscript105.0 KB
scripts/test_fleet_parity.pyscript15.6 KB
scripts/test_grok_bot.pyscript47.7 KB
scripts/test_ground_truth_recording.pyscript68.8 KB
scripts/test_herdr.pyscript14.1 KB
scripts/test_host_policy.pyscript24.6 KB
scripts/test_launch_values.pyscript8.0 KB
scripts/test_metered_usd.pyscript60.4 KB
scripts/test_model_resolution.pyscript9.2 KB
scripts/test_named_route_facts.pyscript28.6 KB
scripts/test_native_hosts.pyscript6.0 KB
scripts/test_new_user_fleet.pyscript3.3 KB
scripts/test_no_hardcode_classification.pyscript75.1 KB
scripts/test_no_legacy_scoring.pyscript6.5 KB
scripts/test_override_store.pyscript2.3 KB
scripts/test_plan_says_why.pyscript44.9 KB
scripts/test_provider_inventory.pyscript83.2 KB
scripts/test_provider_names.pyscript126.0 KB
scripts/test_provider_pools.pyscript14.2 KB
scripts/test_public_data.pyscript93.5 KB
scripts/test_quota_exhaustion.pyscript5.6 KB
scripts/test_rate_limit_windows.pyscript6.3 KB
scripts/test_record_says_why.pyscript26.5 KB
scripts/test_reference_provenance.pyscript16.7 KB
scripts/test_refresh_entry.pyscript110.7 KB
scripts/test_released_names.pyscript4.1 KB
scripts/test_reset_priority.pyscript4.8 KB
scripts/test_roll_answers.pyscript44.7 KB
scripts/test_roll_notifications.pyscript10.4 KB
scripts/test_route_judge.pyscript99.8 KB
scripts/test_route_simulation.pyscript7.0 KB
scripts/test_routing_parity.pyscript32.3 KB
scripts/test_routing_review.pyscript25.1 KB
scripts/test_setup_discovery_parity.pyscript9.5 KB
scripts/test_stage_help.pyscript2.0 KB
scripts/test_state_upgrade.pyscript13.8 KB
scripts/test_turn_refusal.pyscript16.7 KB
scripts/test_usage_cache.pyscript4.9 KB
scripts/test_usage_collector.pyscript21.4 KB
scripts/test_usage_core_claude_accounts.pyscript1.8 KB
scripts/test_usage_core_grok_cli_billing.pyscript15.6 KB
scripts/test_usage_core_magpie.pyscript5.4 KB
scripts/test_usage_core_magpie_pool.pyscript10.8 KB
scripts/test_usage_core_safety.pyscript24.4 KB
scripts/test_user_scores.pyscript102.3 KB
scripts/test_venice_usage.pyscript1.3 KB
scripts/test_window_scope.pyscript5.5 KB
scripts/turn_refusal.pyscript10.7 KB
scripts/usage_cache.pyscript4.7 KB
scripts/usage_collector.pyscript7.4 KB
scripts/usage_core.pyscript142.0 KB
scripts/usage_magpie.pyscript10.6 KB
scripts/usage_reading.pyscript17.1 KB
scripts/user_scores.pyscript72.8 KB
scripts/venice_usage.pyscript2.6 KB
scripts/window_scope.pyscript3.2 KB
scripts/worker_context.pyscript8.9 KB
scripts/worker_sessions.pyscript10.9 KB

calibration/series.py

View on GitHub
"""Same-series siblings and the base a converted model takes (round 058, user 2026-11-10: "后续同系列新出,缺少实测数据的情况下,
先按照同系列的全分数进行折算,不要留下任何一个默认值或者fallback"; five_axis.py estimates through the tier on top of it).

Relatives are declared (owner 2026-11-19, round 6 item 4), never read off a name or a member list: every `families` row
declares its `maker`, its `product` (the maker's product name the vendor releases its lines under: GPT, gpt-oss,
Gemini, Gemma, Claude, ...; round 7), its `line` within the product and its `size` (`small` or `full`), and a row takes
those of the `families` row that decides its name (model_resolution.family_row; `identity`); a `same_model` row takes its
declared model's (it is that model). A row's relatives for an **ability** facet (`groups(acc, 'ability')`) are the rows of
the same maker, product and size, every line of the product (an open-weights product of the same maker is another
product, never a line of this one), a row of another line only up to the declared version distance
(`estimation_order.sibling_first_up_to_distance`); for a **speed or price** facet (`groups(acc, 'line')`: responsiveness, Time/Task, cost per
hour) only those of its own line (a Pro model does not have Astra's speed). A row whose deciding row declares no maker has
no relatives.

The downward relation (`tier_up`, owner 2026-11-18) is a second relation, not series membership: a small row listed in
its `families` row's `models` (a current member of that family) whose `families` row declares `tier_up` (the pattern of
the same line's next larger family row) has, as its larger model, the first model of that larger row the catalog has;
`tier_up_version` declares whether that model is the same version as the row's current members (`same`) or the nearest
version of the line (`nearest`). An older member of the line that is no current member gets its larger model only from
the row's `tier_up_older` declaration ({model: {model: larger, version}}). five_axis.py uses it only for a facet the small
row's own same-size series and the tier fit leave without any value: the small row's own gap only for `same`, the fleet's
size step for both.

A `same_model` row (accounts.json) declares that one row is the very model another row names (a vendor statement with its
url): the declared model is that row's base before any other sibling, and the row carries the type scores of the
categories the row lists under `carries`; a category it does not list is the channel's own and stays missing (a
property of the channel, not of the model behind it).

A `between` row (accounts.json) is two models at once, a planner and an implementer (Antigravity's Orchestrator options):
five_axis.py reads it at a declared share between them. It is no member of a line or a size, so it takes no identity
(`identity` leaves it out, whatever family pattern its name matches): it has no relatives and is no one's relative.

The base for a score is the relative that carries a direct value for it, nearest by the declared version distance
(accounts.json `series_versions`: the rank difference among the distinct declared versions of the row and its relatives,
0 the same version; a row without a declared version ranks after every declared distance), then the same line before
another line at equal distance; relatives still tied form one group whose estimates the caller takes the median of. A
name never decides (round 5 review D: an alphabetical tie-break moved estimates by whole levels when a row was added).
"""
import sys

import sources

sys.path.insert(0, str(sources.REPO / 'scripts'))
import model_resolution  # noqa: E402


def small_rows(acc):
    """The rows of `acc` whose deciding `families` row declares `size: small`."""
    fams = model_resolution.family_patterns(acc)
    return {k for k, v in (acc.get('models') or {}).items()
            if isinstance(v, dict) and (model_resolution.family_row(k, fams) or {}).get('size') == 'small'}


FAMILY_KEYS = ('maker', 'product', 'line', 'size')


def identity(acc):
    """{row: {'maker', 'product', 'line', 'size'}} (module docstring): the declaration of the `families` row that decides
    the row's name, or, for a `same_model` row, its declared model's; a row no `families` row decides and a `between`
    row are left out. A
    `families` row without a nonempty `maker`, `product` and `line` and a `size` of `small` or `full` stops the build."""
    models = {k: v for k, v in (acc.get('models') or {}).items() if isinstance(v, dict)}
    fams = model_resolution.family_patterns(acc)
    for f in fams:
        if not (all(isinstance(f.get(k), str) and f[k].strip() for k in ('maker', 'product', 'line'))
                and f.get('size') in model_resolution.FAMILY_SIZES):
            raise ValueError(f"accounts.json families row {f.get('pattern')!r} must declare maker, product, line and size "
                             '(small or full)')
    out, two = {}, set(between(acc))
    for k in models:
        if k in two:
            continue        # a `between` row is two models, of no one line or size (module docstring)
        row = model_resolution.family_row(k, fams)
        if row is not None:
            out[k] = {key: row[key] for key in FAMILY_KEYS}
    for row, entry in same_model(acc).items():
        if row in models and entry.get('model') in out:
            out[row] = dict(out[entry['model']])
    return out


def groups(acc, scope='ability'):
    """{row key: sorted relative keys} over the score rows of `acc` (module docstring): `ability` the same maker, product
    and size, `line` also the same line."""
    if scope not in ('ability', 'line'):
        raise ValueError(f'scope {scope!r}: ability or line')
    ident = identity(acc)
    rows = [k for k, v in (acc.get('models') or {}).items() if isinstance(v, dict)]

    def key(k):
        i = ident[k]
        return (i['maker'], i['product'], i['size']) + ((i['line'],) if scope == 'line' else ())
    return {k: sorted(j for j in ident if j != k and key(j) == key(k)) if k in ident else [] for k in rows}


def lines(acc):
    """{row: its declared line} (`identity`)."""
    return {k: v['line'] for k, v in identity(acc).items()}


def tier_up(acc):
    """{small row: {'model': its larger model, 'version': 'same' | 'nearest'}} (module docstring, the downward relation)."""
    models = {k: v for k, v in (acc.get('models') or {}).items() if isinstance(v, dict)}
    fams = model_resolution.family_patterns(acc)
    by_pattern = {f['pattern']: f for f in fams}
    out = {}
    for k in sorted(models):
        row = model_resolution.family_row(k, fams) or {}
        if row.get('size') != 'small' or not row.get('tier_up') or k not in row['models']:
            continue
        larger = next((t for t in by_pattern[row['tier_up']]['models'] if t in models), None)
        if larger:
            out[k] = {'model': larger, 'version': row['tier_up_version']}
    for row in fams:
        for k, rel in (row.get('tier_up_older') or {}).items():
            if k in models and rel['model'] in models:
                out[k] = {'model': rel['model'], 'version': rel['version']}
    return dict(sorted(out.items()))


def same_model(acc):
    """{row key: {'model', 'carries', 'statement', 'url', 'as_of', ...}}: the `same_model` rows of accounts.json."""
    return {k: v for k, v in (acc.get('same_model') or {}).items() if isinstance(v, dict)}


def between(acc):
    """{row key: {'planner', 'implementer', 'share', ...}}: the `between` rows of accounts.json (five_axis.between_declaration
    validates them)."""
    return {k: v for k, v in (acc.get('between') or {}).items() if isinstance(v, dict)}


def boards(meas):
    """{model: {source: value}}: the numeric cited readings on the five-axis component dimensions of
    public-measurements.json (first entry per source)."""
    dims = (meas or {}).get('axes') or {}
    out = {}
    for m, fields in ((meas or {}).get('models') or {}).items():
        got = out.setdefault(m, {})
        for dim in dims:
            for e in fields.get(dim) or ():
                v = e.get('value')
                if e.get('source') and e['source'] not in got and isinstance(v, (int, float)) and not isinstance(v, bool):
                    got[e['source']] = v
    return out


def versions(acc):
    """{row: declared version number} (accounts.json `series_versions`)."""
    return dict(((acc.get('series_versions') or {}).get('versions') or {}))


def distance(target, other, series_rows, declared):
    """The version distance of `other` from `target` within the series `series_rows` (target's siblings): the steps
    between their ranks among the series' distinct declared versions; None when either has no declared version."""
    if target not in declared or other not in declared:
        return None
    ladder = sorted({declared[x] for x in (target, *series_rows) if x in declared})
    return abs(ladder.index(declared[target]) - ladder.index(declared[other]))


def nearest(target, candidates, series_rows, declared, small, same=None, line_of=None, other_line_up_to=None):
    """[(version distance, [tied candidates])], nearest first (see the module docstring). `series_rows`: the target's
    relatives; `declared`: `versions(acc)`; `small`: `small_rows(acc)`; `same`: `same_model(acc)`, whose declared model
    for `target` comes first, alone; `line_of`: `lines(acc)`, the same line first at equal distance;
    `other_line_up_to`: a candidate of another line is used only up to that version distance (an ability facet's
    relatives of every line, round 6 item 4). Candidates with one key form one group, in no order a name decides."""
    model = ((same or {}).get(target) or {}).get('model')
    line_of = line_of or {}
    groups = {}
    for c in candidates:
        d = distance(target, c, series_rows, declared)
        other_line = int(line_of.get(c) != line_of.get(target))
        if other_line and c != model and other_line_up_to is not None and (d is None or d > other_line_up_to):
            continue
        k = ((-1, 0, 0, 0) if c == model else
             (0 if d is not None else 1, d or 0, other_line, int((c in small) != (target in small))))
        groups.setdefault(k, []).append(c)
    return [(None if k[0] > 0 else 0 if k[0] < 0 else k[1], sorted(v)) for k, v in sorted(groups.items())]
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall