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.
git clone --depth 1 https://github.com/alifanov/darkflowWrote 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/commands/alifanov/darkflow/fix-issues)<a href="https://agentmods.dev/commands/alifanov/darkflow/fix-issues"><img src="https://agentmods.dev/badge/commands/alifanov/darkflow/fix-issues/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/commands/alifanov/darkflow/fix-issues"><img src="https://agentmods.dev/badge/commands/alifanov/darkflow/fix-issues.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 | $0.00014 | $0.02872 |
| Opus 5.5 | $0.00006 | $0.01149 |
| Sonnet 5.5 | $0.00003 | $0.00574 |
| Haiku 4.5 | $0.00001 | $0.00287 |
Grade A, and why
fix-issues scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Step 1 — Read project config
Load the project config (contract in .darkflow.d/claude.md → Project config). Uses: branch, mergeStrategy, language.
Step 2 — Pick the next task
Pick exactly one open task with status:approved, choosing strictly by priority. The priority order, highest first, is:
priority:criticalpriority:highpriority:mediumpriority:lowstatus:approvedwith no priority set (treat as lowest)
Within a level, take the oldest task (smallest number).
Skip tasks that are not actually actionable here, even if they still carry status:approved:
action:reply— handled exclusively bymailbox-check.scheduledForin the future — the task is snoozed ("don't pick up before this date"); it becomes eligible automatically once the date passes.
Rank every selectable task and take the single best one — pipe df task list's JSON through jq:
n=$(~/.darkflow/df task list --status approved | jq -r '
def prio($p):
if $p == "critical" then 0
elif $p == "high" then 1
elif $p == "medium" then 2
elif $p == "low" then 3
else 4 end;
[ .[] | select(.action != "reply"
and (.scheduledFor == null or .scheduledFor <= (now | todate)))
| {number, rank: prio(.priority)} ]
| sort_by([.rank, .number]) | .[0].number // empty')
If $n is empty, stop — skip the run.
Step 3 — Read the task
Fetch the full task content before touching any code:
~/.darkflow/df task view $n
Read the title, body, and all comments carefully. If the task references other tasks, read those too.
The body carries a ## Plan — it is a starting point written when the finding was filed, not a
contract. Follow it where it still holds; where the code has moved on, rewrite it and say so in
the closing comment. The ## Acceptance criteria are the contract: each one names the command,
report or URL that decides it, and every one must pass before the task closes.
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 First seen · 220 lines · 14 tokens per session scan A a654478b789e
fix-issues is a command published in the GitHub repository alifanov/darkflow (4 stars, last pushed yesterday), licensed MIT. It adds 14 tokens to every session and 2,872 once invoked, about $0.0001 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 commands, from other repositories
issue
Command "issue" from getsentry/toolkit, covering examples, list issues, list issues in a specific project, several projects in the same org and all projects in an org.
local
Command "local" from getsentry/toolkit, covering examples, start the server and tail events (default), run your app with the local server auto-enabled, use a custom port and only show errors and logs (filter out transactions).
debug-files
Command "debug-files" from getsentry/toolkit, covering examples, inspect a debug information file (auto-detects the format), bundle jvm sources with a debug id, exclude additional directories and output as json.
feedback
Modern User Feedback is stored as Feedback issues. The command always limits issue searches to issue.category:feedback; it does not use the legacy User Reports API.
event
Command "event" from getsentry/toolkit, covering examples, sending events, send an error event (default level), specify level, release, and environment and add tags and extra data.
log
Command "log" from getsentry/toolkit, covering examples, list logs, list last 100 logs (default), show only error logs and filter by message content.