Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add aahlijia/dekko --skill dekko-daemongit clone --depth 1 https://github.com/aahlijia/dekkoWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/aahlijia/dekko/dekko-daemon)<a href="https://agentmods.dev/skills/aahlijia/dekko/dekko-daemon"><img src="https://agentmods.dev/badge/skills/aahlijia/dekko/dekko-daemon/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/aahlijia/dekko/dekko-daemon"><img src="https://agentmods.dev/badge/skills/aahlijia/dekko/dekko-daemon.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00094 | $0.01012 |
| Opus 5.5 | $0.00038 | $0.00405 |
| Sonnet 5.5 | $0.00019 | $0.00202 |
| Haiku 4.5 | $0.00009 | $0.00101 |
Grade A, and why
dekko-daemon scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Warming up repeated dekko CLI calls with the daemon
dekko daemon start spawns a small per-repo background process that
keeps a warm, in-memory map index across CLI invocations, instead of
each one reloading and re-indexing map.json from disk. It only
matters for the bare CLI path (commands run via Bash) — if you're
calling the MCP tools (mcp__dekko__*) instead, skip this: the MCP
server already keeps its own warm per-process cache
(Context.index_cache) independently, by design, so nothing here
applies there.
When to start it
dekko daemon start # spawns it, returns immediately
Worth doing at the start of a Bash-CLI-heavy stretch of work — a
refactor loop running query/affected/workset repeatedly, a batch
of outline/search calls while orienting in a large repo, or
scripted/hook invocations. Skip it for a one-off dekko map or a
single query; the daemon adds spin-up overhead that isn't worth it for
one call. On a small repo (fast reload already), it's rarely worth
the trouble either way.
Read-only subcommands route through it transparently once running
(query, search, workset, diff, affected, outline,
context, trace, stats, summary, lean, unused, ambiguous,
deps, status, ledger, note list, export) — same output and
exit code, just without the reload. Write-path commands (map, note add/note rm, hooks ...)
always run directly regardless. Pass --no-daemon on any command to
force direct execution for that one call.
Checking it and shutting down
dekko daemon status # running? pid, uptime, cache hits/misses
dekko daemon stop # graceful shutdown
No need to proactively stop it — it self-shuts-down after 30 minutes idle by default. Only stop it explicitly if you need a guaranteed-cold next read (rare) or are done with a long session and want to be tidy.
What it does not speed up
diff/affected's old-side (the git rev being diffed against) reparse is not covered by the warm cache at all — that cache only ever holds the current working tree's index. A comparison against a rev it hasn't seen before pays the same reparse cost a direct invocation would; only a repeat comparison against the same already-cached rev benefits (via the separate on-disk.dekko/rev-cache/, shared with direct calls regardless of daemon use).- It never trades correctness for speed: every read re-validates the
cached index the same way a direct invocation would, so a
working-tree edit or an out-of-band
dekko mapis never served stale.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed 05117d9412b6
- 2mo ago First seen · 86 lines · 94 tokens per session scan A 0ac7f6c7c5b8
dekko-daemon is a skill published in the GitHub repository aahlijia/dekko (6 stars, last pushed 15d ago), licensed MIT. It adds 94 tokens to every session and 1,012 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
roam
Codebase comprehension via roam-code CLI. Use when exploring codebases, planning modifications, debugging failures, assessing PR risk, or checking architecture health. Triggers on: understanding project structure, pre-change safety checks, finding symbols/files, blast radius analysis, affected tests, health scoring…
release-tag
Cut and publish a codeweb release end-to-end — prep the version/changelog file ecosystem, land it on main, push the vX.Y.Z tag, and let the tag-triggered workflow publish the GitHub Release. Use when asked to "cut a release", "tag a release", "publish vX.Y.Z", "do the release", or "ship version X".
codeweb
Inspect mapped callers and impact before changing code, find existing implementations before adding one, and review structural changes after an edit. Use during coding tasks; Codeweb product planning uses a separate workflow.
plan-schema
Plan file (.claude/plans/-plan.md) reference — sections, per-file format, dependency graph, and converter mappings.
beads-schema
Beads (bd CLI) reference — issue types, statuses, priorities, dependencies, and common commands.
prd-schema
Skill "prd-schema" from GantisStorm/essentials-claude-code, covering when to use, conversion from plans, schema, root object and user story object.