moadim 0.19.1

Loop engine for AI agents — routines over REST, MCP, and a built-in web UI
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

# Configuration for the `lychee` Markdown link checker run by
# .github/workflows/lychee.yml (issue #431). See
# https://lychee.cli.rs/usage/config/ for the full schema.
#
# The gate is meant to be signal, not noise: known-flaky/placeholder hosts are
# ignored outright, and transient failures get retried before a link is
# declared dead, so only durably-broken links fail CI.

# Directories that hold build artifacts or dependencies, not hand-written
# docs — scanning them only wastes CI time and can false-positive on
# generated content.
exclude_path = [
  "target/",
  "node_modules/",
]

# Loopback dev-server URL documented in README.md's "Bind address" section
# (`http://localhost:5784/docs`); nothing listens on it outside a running
# local server, so it can never resolve in CI.
exclude_loopback = true

# Placeholder domains used in examples/config snippets, never meant to
# resolve to a real host.
exclude = [
  '^https?://(.*\.)?example\.(com|org|net)',
  # trunkrs.dev (linked from CONTRIBUTING.md's Prerequisites) sits behind a
  # bot wall that 403s automated/CI requests (including a real browser UA
  # from this network) regardless of reachability — a known false positive,
  # not a dead link. Verified manually before adding; revisit if the site
  # changes its bot-protection setup.
  '^https?://(www\.)?trunkrs\.dev',
]

# Retry/backoff so a transient network blip or upstream rate limit doesn't
# fail the gate — only a durably-dead link should.
max_retries = 3
retry_wait_time = 5
timeout = 20

# Some hosts (e.g. shields.io, crates.io) reject bare HEAD requests, so use
# GET directly rather than false-positiving on those.
method = "get"

# Treat "too many requests" as acceptable rather than a broken link — it
# means the host is alive, just rate-limiting us.
accept = ["200..=299", "429"]