report
Skill0 starsChanged 3 months agoMITView on GitHub
File a report or piece of feedback from any Hermes platform session. Load this skill when the user says "/report", "file a bug", "report a bug", "something's broken", "something's wrong", "that's weird", "this is wrong", "this isn't working", "log this issue", or any other signal that they want to capture a problem for triage. Works on Telegram, Discord, Slack, and the CLI — the closed loop (a notification back to this chat when the issue is resolved) is wired automatically for gateway sessions.
- Reads credentials
104 other projects use a version of this.
Adopted by
Most recently taken up by
- flegaspi700/pmo-project-standardizationtheir own version, first seen Nov 16, 2026
- TsukiNi22/skillstheir own version, first seen Nov 16, 2026
- TsukiNi22/ai-utilstheir own version, first seen Nov 16, 2026
- Shugar86/methodist-agenttheir own version, first seen Nov 16, 2026
- arxdsilva/ai-loopstheir own version, first seen Nov 16, 2026
- wildblue-ai/llm-secure-patternstheir own version, first seen Nov 16, 2026
Used with
Projects with it also use opportunities, export and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
ansvisorSkillopportunities
Finds content gaps, prompts worth tracking, and the off-site sources AI cites instead of the brand.
Also in 3 projects that use this
soilmassSkillexport
Animation export configurations including format selection, resolution, framerate, quality, and optimization settings. Use when rendering animations for web, mobile, video, or social media. Provides tokens for all export parameters across Lottie, GIF, video, and sprite sheet formats.
Also in 4 projects that use this
GoogleSkillquality
Evaluates whether a GitHub issue is spam, empty, needs more information, or is OK to proceed.
Also in 3 projects that use this
genomewalkerSkillresume
Resume a thread by loading its ~800-token context capsule
Also in 7 projects that use this
tokyubevoxelverseSkillcheckup
Give a repository a full health exam — doc rot, dependency staleness, CI decay, git hygiene, dead links, docs contradicting code — and grade it. Use when the user wants to know what quietly rotted, or a prioritized maintenance list.
Also in 3 projects that use this
rajnavakotiSkillscan
Run the AI dashboard scanner and generate the HTML report.
Also in 5 projects that use this
Other versions104 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 104
- microsoft/data-formulator★ 17ksame name
- Zafer-Liu/Data-Analysis-Agent★ 2.6ksame name
- H-mmer/pentest-agents★ 815same name
- cwinvestments/memstack★ 433same name
- ai-search-guru/getcito-worlds-first-open-source-aio-aeo-or-geo-tool★ 432same name
- qbeka/jev-job-search★ 368same name
- automagik-dev/genie★ 353same name
- automagik-si/genie★ 353same name
- namastexlabs/automagik-genie★ 351same name
- ansvisor/ansvisor★ 241same name
94 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository36 other files it gives its agents.
- check-upstream-firstSkill
- data-verificationSkill
- deep-diveSkill
- delegation-handoffSkill
- grok-searchSkill
- is-it-really-brokenSkill
- keep-goingSkill
- learn-from-conversationsSkill
- memory-cleanupSkill
- moa-solveSkill
- mob-checkSkill
- multi-reviewSkill
- project-stewardSkill
- prove-the-setting-worksSkill
- recallSkill
- robustify-doctorSkill
- skill-librarianSkill
- trust-frameworkSkill
- address-pr-commentsSkill
- diagram-renderingSkill
- find-the-real-bottleneckSkill
- live-database-maintenanceSkill
- mini-appSkill
- pr-review-sweepSkill
- scheduled-job-runnerSkill
- stop-the-noiseSkill
- upgrade-a-live-service-safelySkill
- vendor-switch-checkSkill
- writing-for-non-engineersSkill
- email-stewardSkill
- fieldySkill
- google-docsSkill
- google-sheetsSkill
- google-slidesSkill
- imessage-bluebubblesSkill
- vapi-callsSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. 0 are common, 5 are distinctive, and 8 are very specific to this project.
0
Common
Most of the 76 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 47
Gather the title: If the user's message contains a clear description, distill it into ≤10 words.
- Distinctive: line 56
Do NOT include raw API keys, tokens, or personal credentials anywhere in the body.
- Very specific: line 58
File the report: Prefer the helper script if it is installed with this skill (see Script Path below).
- Distinctive: line 68
Never drop the report: If filing fails, save the markdown body to a local dropfile and tell the maintainer where it is.
- Very specific: line 89
If the script is not found, fall back to calling hermes kanban create inline (see Fallback section below).
- Very specific: line 150
Run this with TITLE, BODY_FILE, and REPORTER exported.
- Very specific: line 212
After a successful ok: true result:
- Distinctive: line 217
After a deduped result (same session within the 2-minute window, same id returned):
- Distinctive: line 221
After a failure with dropfile:
- Very specific: line 247
If REPORT_BOARD_OWNER is set and matches this profile → local.
- Very specific: line 255
When the triager calls hermes kanban complete <id>, the gateway's kanban notifier automatically delivers one message to the subscribed …
- Very specific: line 262
If the card is blocked: ⏸ Kanban...
- Distinctive: line 271
Do not include personal credentials, API keys, or sensitive financial/health context in the body.
“Common” is measured: kept by at least 6 in 10 of the 76 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Report
File a bug or piece of feedback, land it in the triage column of the shared kanban
board, and wire a closed-loop notification so the reporter hears back when it's resolved
— all from a single /report or a natural-language sentence.
When to load
Load this skill automatically when the current user turn contains any of these patterns (case-insensitive):
/report(exact slash-command)- "file a bug" / "report a bug" / "report this bug"
- "file a report" / "file an issue" / "log this issue"
- "something's broken" / "something is broken"
- "something's wrong" / "something is wrong"
- "this is broken" / "this isn't working" / "this doesn't work"
- "that's weird" / "this is weird" / "that was weird"
- "there's a bug" / "i found a bug"
Also load when the user explicitly says "load the report skill" or asks to file feedback or a suggestion.
What you do
- Acknowledge immediately. Reply with a short line so the user knows you've got it:
"On it — filing that now." - Gather the title. If the user's message contains a clear description, distill it
into ≤10 words. If they just said
/reportwith no detail, ask one brief question:"What's the issue in one sentence?". Do not ask for more — the transcript excerpt provides context. - Build the body. Compose a markdown body with:
- What the user reported: their exact words (quote them)
- Session context: profile name, platform, timestamp (from env)
- Transcript excerpt: the last 5–10 turns of the current session in condensed form (enough to let a triager reproduce or understand)
- Do NOT include raw API keys, tokens, or personal credentials anywhere in the body.
- Write body to a temp file (e.g.
/tmp/report-body-<session_id>.md). - File the report. Prefer the helper script if it is installed with this skill (see
Script Path below). If only this
SKILL.mdis installed, use the self-contained fallback workflow below — do not fail just becausescripts/file_report.pyis absent. - Confirm.
- Local path: tell the user the card id and: "You'll get a ping here when it's resolved."
- Remote path: tell the user the report was sent to the triage board and they will get a ping when it is triaged or resolved. The remote path is asynchronous, so it may not have a card id immediately.
- Never drop the report. If filing fails, save the markdown body to a local dropfile and tell the maintainer where it is.
Script path
The helper lives alongside this skill. Resolve it relative to this file:
SKILL_DIR=$(dirname "$(hermes skills list --paths 2>/dev/null | grep '/report/SKILL.md' | head -1)")
SCRIPT="$SKILL_DIR/scripts/file_report.py"Or resolve via Hermes's skills directory convention:
# The skill is installed at one of these locations:
# ~/.hermes/skills/report/scripts/file_report.py (single-profile install)
# ~/.hermes/profiles/<name>/skills/*/report/scripts/file_report.py (profile install)
SCRIPT=$(find ~/.hermes -path "*/report/scripts/file_report.py" 2>/dev/null | head -1)If the script is not found, fall back to calling hermes kanban create inline (see
Fallback section below).
Invocation
python3 "$SCRIPT" \
--title "<10-word title you distilled>" \
--body-file /tmp/report-body-<session_id>.md \
--reporter "<user's display name or Telegram handle>" \
--jsonParse stdout as JSON:
{
"ok": true,
"path": "local",
"task_id": "t_a1b2c3d4",
"status": "triage",
"subscribed": true
}ok: true→ confirm to the userok: false, dropfile: "..."→ post that path as a note and DM the maintainerok: false, error: "..."→ use the fallback path below
Fallback (if script not found)
The fallback is self-contained so a single-file skill install still works.
Migration note: if this host still has pre-rename
BUG_WEBHOOK_URL/BUG_BOARD_OWNERset but not the newREPORT_*names, do NOT fall through to the local path — that would drop the report off the shared queue. Rename the env vars to theirREPORT_*equivalents first, then file.
Local fallback (no REPORT_WEBHOOK_URL configured)
hermes kanban create "<title>" \
--triage \
--tenant fleet-reports \
--created-by "<reporter>" \
--body "<condensed body>" \
--jsonThen subscribe the reporter's current chat, using fields that exist in the session env:
hermes kanban notify-subscribe <task_id> \
--platform <HERMES_SESSION_PLATFORM> \
--chat-id <HERMES_SESSION_CHAT_ID> \
[--thread-id <HERMES_SESSION_THREAD_ID>] \
[--user-id <HERMES_SESSION_USER_ID>]Remote fallback (REPORT_WEBHOOK_URL configured)
Run this with TITLE, BODY_FILE, and REPORTER exported. It posts to the board owner
using Hermes' generic webhook signature contract (X-Webhook-Signature = hex
HMAC-SHA256 of the raw JSON body). The receiver is asynchronous and returns a delivery
id, not a card id.
python3 - <<'PY'
import hashlib, hmac, json, os, pathlib, time, urllib.error, urllib.request
url = os.environ["REPORT_WEBHOOK_URL"]
secret = os.environ["REPORT_WEBHOOK_SECRET"]
title = os.environ["TITLE"]
body = open(os.environ["BODY_FILE"]).read()
reporter_id = (
os.environ.get("HERMES_SESSION_ID")
or os.environ.get("HERMES_SESSION_CHAT_ID")
or os.environ.get("HERMES_PROFILE")
or f"cli-{os.getpid()}"
)
bucket = int(time.time()) // 120
title_slug = hashlib.sha256(title.encode()).hexdigest()[:8]
idem_basis = f"{reporter_id}:{title_slug}:{bucket}"
idem = "report-" + hashlib.sha256(idem_basis.encode()).hexdigest()[:16]
payload = {
"event_type": "report",
"title": title,
"body": body,
"idempotency_key": idem,
"reporter": os.environ.get("REPORTER") or os.environ.get("HERMES_SESSION_USER_NAME", "fleet-user"),
"profile": os.environ.get("HERMES_PROFILE", ""),
"tenant": os.environ.get("REPORT_TENANT") or "fleet-reports",
"platform": os.environ.get("HERMES_SESSION_PLATFORM", ""),
"chat_id": os.environ.get("HERMES_SESSION_CHAT_ID", ""),
"thread_id": os.environ.get("HERMES_SESSION_THREAD_ID", ""),
"user_id": os.environ.get("HERMES_SESSION_USER_ID", ""),
"session_id": os.environ.get("HERMES_SESSION_ID", ""),
"timestamp": int(time.time()),
}
raw = json.dumps(payload, separators=(",", ":")).encode()
sig = hmac.new(secret.encode(), raw, hashlib.sha256).hexdigest()
req = urllib.request.Request(url, data=raw, method="POST", headers={
"content-type": "application/json",
"X-Webhook-Signature": sig,
"X-Idempotency-Key": idem,
})
try:
with urllib.request.urlopen(req, timeout=15) as r:
resp = json.loads(r.read().decode() or "{}")
status = resp.get("status")
if status is not None and status not in {"accepted", "delivered", "duplicate"}:
raise RuntimeError(f"server response not accepted: {resp!r}")
print(json.dumps({"ok": True, "path": "remote", "status": status or "accepted"}))
except (Exception, urllib.error.HTTPError) as e:
drop = pathlib.Path("/tmp") / f"report-{idem}.json"
drop.write_text(json.dumps({"payload": payload, "error": str(e)}, indent=2))
print(json.dumps({"ok": False, "path": "remote", "dropfile": str(drop), "error": str(e)}))
raise SystemExit(3)
PYUser-facing confirmation
After a successful ok: true result:
Filed as t_a1b2c3d4. I'll look into it and you'll get a ping here when it's resolved.
After a deduped result (same session within the 2-minute window, same id returned):
Already filed as t_a1b2c3d4 — no duplicate created.
After a failure with dropfile:
Couldn't reach the board right now, but your report is saved locally and I've flagged it for the maintainer. Nothing dropped.
Configuration
Set these as environment variables in your Hermes profile's dotenv file (the .env next
to your profile config), or export them in the gateway's environment:
| Variable | Required on | Purpose |
REPORT_BOARD_OWNER | All profiles | Profile name that owns the triage board (required — no default ships in the template) |
REPORT_WEBHOOK_URL | Non-owner profiles | Endpoint to POST the report to |
REPORT_WEBHOOK_SECRET | Non-owner profiles | HMAC-SHA256 signing secret |
The board-owner profile (the one where $HERMES_PROFILE matches $REPORT_BOARD_OWNER)
does not need the webhook env vars — it calls hermes kanban directly. All other fleet
members need both webhook vars.
How the script decides local vs remote
Routing is by capability, not by name-matching, so an unconfigured install never silently misroutes and breaks. In priority order:
--force-remote/--force-localflags always win (testing / explicit control).- If
REPORT_BOARD_OWNERis set and matches this profile → local. - Else if
REPORT_WEBHOOK_URLis configured → this is a satellite → remote. - Else (no webhook configured) → local. A single-host install with no webhook files directly to its own board — the safe default (worst case the report lands on this host's board rather than vanishing).
Closed-loop notification
When the triager calls hermes kanban complete <id>, the gateway's kanban notifier
automatically delivers one message to the subscribed chat/thread:
✔ @<user> Kanban t_a1b2c3d4 done — <title>If the card is blocked: ⏸ Kanban... blocked If it fails: ✖ Kanban... gave up
The subscription is removed automatically after the done notification. No polling, no manual DM required.
Privacy and data-handling notes
- Transcript excerpts included in the report body will be visible to the triager (the
maintainer). Only include turns from the session where
/reportwas invoked. - Do not include personal credentials, API keys, or sensitive financial/health context in the body.
- The
fleet-reportskanban board is local (SQLite on the maintainer's machine), not cloud-synced. The webhook payload is HMAC-signed for integrity but transmitted over the network — use Tailscale or another private tunnel, not a public URL.
Triage runbook
See references/triage.md for how the maintainer handles the triage column, closes
cards, and what the automatic notification message looks like.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

