Functions — github.com/FareedKhan-dev/long-running-agent

hub / github.com/FareedKhan-dev/long-running-agent / functions

Functions438 in github.com/FareedKhan-dev/long-running-agent

⨍ Functions 438◇ Types & classes 148

↓ 72 callersMethodget

(self, skill_id: str)

src/lra/memory/skills.py:75

↓ 39 callersMethodappend

(self, event: EventRecord)

src/lra/memory/episodic.py:16

↓ 16 callersFunctiondefault_local_tools

The non-egress tool set (file IO + shell) every mission gets by default.

src/lra/execution/tools/__init__.py:9

↓ 16 callersMethodfailure

(cls, error: str)

src/lra/contracts/tools.py:45

↓ 13 callersFunctionget_settings

Return the process-wide settings (cached).

src/lra/config.py:62

↓ 12 callersMethodcomplete

( self, messages: list[ModelMessage], *, tools: list[dict[str, object]] | None

src/lra/model/stub.py:28

↓ 12 callersMethodopen

Open (or restore from ``snapshot_id``) a session rooted at ``workdir``.

src/lra/contracts/sandbox.py:73

↓ 12 callersMethodrun

Execute the tool with validated ``arguments`` and return a compact result.

src/lra/contracts/tools.py:63

↓ 11 callersMethod_path

(self, name: str)

src/lra/state/mission_anchor.py:44

↓ 11 callersMethodinitialize

(self, *, title: str, description: str, items: Checklist)

src/lra/state/mission_anchor.py:48

↓ 8 callersFunctionbuild_provider

Construct the configured model backend (the only place a concrete LLM is chosen).

src/lra/model/__init__.py:23

↓ 8 callersFunctionis_shared

True if ``path`` is a shared/ambiguous file that must stay on the serial lead thread.

src/lra/coordination/ownership.py:35

↓ 8 callersMethodpermits

Whether ``writer`` may write ``path`` under the single-writer-per-file invariant.

src/lra/coordination/ownership.py:78

↓ 8 callersFunctionrun_git

Run ``git <args>`` in ``cwd`` and return stdout (stripped).

src/lra/state/git_ops.py:18

↓ 7 callersFunction_checklist

()

tests/unit/test_mission_anchor.py:25

↓ 7 callersMethoddispatch

(self, call: ToolCall, ctx: ToolContext)

src/lra/execution/dispatcher.py:40

↓ 6 callersFunction_msgs

()

tests/unit/test_stub_model.py:12

↓ 6 callersMethodembed

(self, texts: list[str])

src/lra/memory/embeddings.py:28

↓ 6 callersMethodobserve

Record a (state, action) signature; return True if it has now looped past threshold.

src/lra/governor/governor.py:66

↓ 6 callersFunctionparse_action

Interpret a model turn as an Action (native tool-calls take precedence over JSON text).

src/lra/agent/loop.py:62

↓ 6 callersMethodpost

Append to the durable main board.

src/lra/coordination/blackboard.py:24

↓ 6 callersMethodread_checklist

Return the full current checklist (all items, not just the open ones).

src/lra/state/mission_anchor.py:54

↓ 6 callersMethodrecord

(self, *, cycle_id: str, usage: Usage, usd: float)

src/lra/governor/cost.py:30

↓ 6 callersMethodrecord

( self, kind: str, *, mission_id: str, cycle_id: str = "", **data: object )

src/lra/obs/events.py:53

↓ 6 callersMethodrun

( self, *, objective: str, ctx: ToolContext, extra_context: str = "" )

src/lra/agents/subagent.py:46

↓ 6 callersMethodsuccess

(cls, content: str = "", *, artifact_ref: str | None = None)

src/lra/contracts/tools.py:41

↓ 5 callersFunction_normalize

(path: str)

src/lra/coordination/ownership.py:31

↓ 5 callersFunction_run

()

src/lra/cli/main.py:107

↓ 5 callersMethodadd

(self, name: str, undo: Callable[[], Awaitable[None]])

src/lra/durable/saga.py:28

↓ 5 callersMethodappend

(self, record: DecisionRecord)

src/lra/coordination/decision_log.py:20

↓ 5 callersMethodauthorize_next

( self, ledger: CostLedger, *, cycles_done: int, projected_usd: float | None = None )

src/lra/governor/governor.py:29

↓ 5 callersMethodclose

Release the session (no-op for local; tears down container/VM for others).

src/lra/contracts/sandbox.py:62

↓ 5 callersFunctionconnect_client

Connect to Temporal with the ClaimCheck data converter installed.

src/lra/durable/worker.py:24

↓ 5 callersMethodencode

(self, payloads: Sequence[Payload])

src/lra/durable/codec.py:35

↓ 5 callersMethodfind

(self, query: str, *, k: int = 3, namespace: str | None = None)

src/lra/memory/skills.py:61

↓ 5 callersMethodplan

(self, *, title: str, description: str, acceptance: str = "")

src/lra/agents/planner.py:59

↓ 5 callersMethodrun

(self, arguments: dict[str, object], ctx: ToolContext)

src/lra/execution/tools/fs.py:108

↓ 5 callersMethodrun_cycle

( self, *, ctx: ToolContext, mission_id: str, cycle_id: str, a

src/lra/agent/loop.py:102

↓ 4 callersFunction_ctx

(tmp_path: Path)

tests/unit/test_tools_dispatcher.py:21

↓ 4 callersMethod_emit

(self, kind: str, mission_id: str, cycle_id: str, **data: object)

src/lra/agent/loop.py:231

↓ 4 callersFunction_records

()

tests/unit/test_hybrid_memory.py:17

↓ 4 callersFunctioncosine

Cosine similarity of two equal-length vectors (0.0 if either is degenerate).

src/lra/memory/semantic_memory.py:16

↓ 4 callersMethoddecode

(self, payloads: Sequence[Payload])

src/lra/durable/codec.py:54

↓ 4 callersFunctionnew_id

Mint a fresh random id like ``mission_3f9a1c0b2d4e``. Activity-only (non-deterministic).

src/lra/ids.py:17

↓ 4 callersFunctionparse_checklist

Extract a checklist from model output; assign 2-digit ids; tolerate messy output.

src/lra/agents/planner.py:24

↓ 4 callersMethodpermits

(self, url: str)

src/lra/safety/egress.py:20

↓ 4 callersMethodquery

(self, text: str, *, k: int = 10)

src/lra/memory/hybrid.py:37

↓ 4 callersMethodread_situational_awareness

(self)

src/lra/state/mission_anchor.py:51

↓ 4 callersFunctionrun_proc

Run ``argv`` in ``cwd`` and return a real ``ExecResult`` (never raises on non-zero exit).

src/lra/execution/proc.py:17

↓ 3 callersFunction_client

(handler: Callable[[httpx.Request], httpx.Response])

tests/unit/test_model_backends.py:20

↓ 3 callersFunction_clip

(text: str)

src/lra/execution/tools/fs.py:19

↓ 3 callersFunction_complete_commits

(workdir: Path)

tests/durability/test_durable_spine.py:47

↓ 3 callersFunction_ctx

(tmp_path: Path)

tests/unit/test_subagent.py:20

↓ 3 callersFunction_init_mission

(tmp_path: Path, n: int, *, cycles_before_can: int = 200)

tests/durability/test_durable_spine.py:36

↓ 3 callersMethodadd

(self, records: list[MemoryRecord])

src/lra/memory/semantic_memory.py:36

↓ 3 callersMethodadd

(self, records: list[MemoryRecord])

src/lra/memory/hybrid.py:32

↓ 3 callersMethodassign

Assign ``path`` to ``writer``. Shared files can only ever belong to the lead.

src/lra/coordination/ownership.py:68

↓ 3 callersFunctionhas_commits

True if the repo has at least one commit.

src/lra/state/git_ops.py:58

↓ 3 callersMethodjudge

(self, *, candidate: str, rubric: str)

src/lra/agents/judge.py:42

↓ 3 callersFunctionmodel_for_role

Return a model tuned to ``role_name``'s tier (Claude backend) or the default model.

src/lra/agents/router.py:16

↓ 3 callersFunctionparse_coverage_pct

Parse the total coverage fraction (0..1) from `pytest --cov` / `coverage` output.

src/lra/verify/trust_bootstrap.py:19

↓ 3 callersFunctionpermits

True if the capability set is safe (holds at most two of the three).

src/lra/safety/rule_of_two.py:27

↓ 3 callersMethodquery

(self, text: str, *, k: int = 5)

src/lra/memory/semantic_memory.py:50

↓ 3 callersMethodread

(self)

src/lra/coordination/blackboard.py:32

↓ 3 callersMethodrecord

(self, *, made_progress: bool)

src/lra/durable/ledgers.py:41

↓ 3 callersFunctionrun_eval_suite

Judge each case and aggregate into a report (pass rate + mean score).

src/lra/evals/harness.py:44

↓ 2 callersFunction_as_int

Coerce a JSON-ish value to a positive int, falling back to ``default``.

src/lra/state/migrations.py:28

↓ 2 callersMethod_deny

(self, reason: str, projected: float, spent: float)

src/lra/governor/governor.py:49

↓ 2 callersMethod_ensure_pool

(self)

src/lra/memory/semantic_pg.py:28

↓ 2 callersFunction_execute_cycle

Advance the mission by one verified item via the real agent loop; idempotent.

src/lra/durable/activities.py:26

↓ 2 callersFunction_ignored

(path: Path)

src/lra/execution/tools/fs.py:15

↓ 2 callersMethod_permitted

(self, name: str)

src/lra/execution/dispatcher.py:34

↓ 2 callersFunction_setup

(tmp_path: Path, items: Checklist)

tests/unit/test_agent_loop.py:32

↓ 2 callersFunction_tokenize

(text: str)

src/lra/memory/hybrid.py:19

↓ 2 callersMethod_write_checklist

(self, checklist: Checklist)

src/lra/state/mission_anchor.py:119

↓ 2 callersMethodadd

(self, skill: Skill)

src/lra/memory/skills.py:44

↓ 2 callersMethodappend_event

(self, event: EventRecord)

src/lra/state/mission_anchor.py:58

↓ 2 callersFunctioncheck_rule_of_two

Raise ``RuleOfTwoViolation`` if a session would hold the full lethal trifecta.

src/lra/safety/rule_of_two.py:32

↓ 2 callersMethodcommit_checkpoint

(self, checkpoint: Checkpoint)

src/lra/state/mission_anchor.py:61

↓ 2 callersFunctioncompact_messages

Return a compacted message list (system + summary-of-older + last ``keep_last`` turns).

src/lra/agent/compaction.py:19

↓ 2 callersMethodcompensate

Run compensations newest-first; return the names that ran. Records any undo failures.

src/lra/durable/saga.py:31

↓ 2 callersMethodcomplete

( self, messages: list[ModelMessage], *, tools: list[dict[str, object]] | None

src/lra/model/failover.py:22

↓ 2 callersMethodcomplete

( self, messages: list[ModelMessage], *, tools: list[dict[str, object]] | None

src/lra/model/openai_compat.py:39

↓ 2 callersMethodcomplete

( self, messages: list[ModelMessage], *, tools: list[dict[str, object]] | None

tests/unit/test_more_modules.py:19

↓ 2 callersFunctionconfigure_logging

Configure process-wide structlog output (console for dev, JSON for prod/ingest).

src/lra/obs/events.py:17

↓ 2 callersFunctionconsolidate

Distill ``episodes`` into semantic ``MemoryRecord`` facts (model-driven, safe fallback).

src/lra/memory/consolidation.py:37

↓ 2 callersFunctiondecide_safe_park

Park iff a CRITICAL dependency is DOWN; otherwise continue (noting degraded optionals).

src/lra/ops/degradation.py:39

↓ 2 callersMethoddispatch

Validate + execute a tool call (never raises; failures come back as ``ToolResult``).

src/lra/contracts/tools.py:76

↓ 2 callersMethodestimate_cost_usd

(self, usage: Usage)

src/lra/model/stub.py:55

↓ 2 callersMethodexec

Run ``argv`` (no shell) and capture its real output + exit code.

src/lra/contracts/sandbox.py:43

↓ 2 callersMethodmean_usd_per_cycle

(self)

src/lra/governor/cost.py:53

↓ 2 callersMethodopen

(self, *, workdir: str, snapshot_id: str | None = None)

src/lra/execution/sandbox_local.py:59

↓ 2 callersMethodput

(self, data: bytes)

src/lra/persistence/object_store.py:33

↓ 2 callersMethodquery

(self, text: str, *, k: int = 5)

src/lra/contracts/memory.py:58

↓ 2 callersMethodread_responses

(self)

src/lra/coordination/blackboard.py:35

↓ 2 callersMethodread_situational_awareness

Reconstruct full mission awareness from git + the anchor files (seconds, any restart).

src/lra/contracts/state.py:113

↓ 2 callersFunctionreciprocal_rank_fusion

Fuse ranked id-lists via RRF: score = sum 1/(k + rank).

src/lra/memory/hybrid.py:66

↓ 2 callersFunctionreflect_on_failure

Produce a short post-mortem for a failed attempt (no code, just lessons).

src/lra/agents/reflection.py:15

↓ 2 callersFunctionrender_tools

(specs: list[ToolSpec])

src/lra/agent/prompt.py:23

↓ 2 callersMethodrequest

(self, req: GateRequest)

src/lra/hitl/gate.py:24

next →1–100 of 438, ranked by callers