symphony-trello
AGENTS.md5 starsChanged 15 days agoApache-2.0View on GitHub
These instructions are repository-local and apply to Codex or any other AI agent working in this checkout. Follow them in addition to higher-priority system, developer, and user instructions. Symphony for Trello is a senior-engineered Quarkus backend on Java 25 LTS and Maven 3 that implements the adapted Symphony-for-Trello specification in SPEC.md; work as a senior Quarkus engineer driven by that spec. Package manager: Maven via ./mvnw (do not invoke a system Maven). The few Node-based tools use pnpm at the…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,542 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository3 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- live-bugbashSkill
- publish-bugbash-issuesSkill
The file
We split this AGENTS.md into the rules it gives an agent
12 rules. About 0 are common, 11 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈11
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Before creating any commit that may be pushed, verify the configured author against the authenticated GitHub account even when the checkout …
- Very specific: line 16
Before every push, inspect the author of every outgoing commit.
- Distinctive: line 21
Every commit that reaches main is one cohesive change, on every pull request, outside contributors' included.
- Distinctive: line 25
Commit History in Main checkboxes after every push.
- Distinctive: line 29
Before starting any task, open docs/agents/default-workflow.md and follow it.
- Distinctive: line 32
When a user corrects a repeatable mistake or states a generally useful working preference, make the durable agent-doc update proactively in …
- Distinctive: line 38
Add a regression test or another durable guardrail proportionate to the failure.
- Distinctive: line 41
When an installed skill eval issue captures bad or missing behavior, add or update the corresponding temporary repo-local override …
- Distinctive: line 52
When adding or editing tests, apply the AssertJ and test-duplication guidance in Testing, including avoiding assertion loops or assertion …
- Distinctive: line 56
Follow the compatibility and migration rules in Compatibility discipline.
- Distinctive: line 59
Never let a GitHub Actions workflow run or restore anything a pull request controls in a privileged context.
- Distinctive: line 67
When asked to "rebase main to the Symphony repo" or similar, follow the upstream Symphony rebase process in Default workflow.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Agent Operating Instructions
These instructions are repository-local and apply to Codex or any other AI agent working in this checkout. Follow them in addition to higher-priority system, developer, and user instructions.
Symphony for Trello is a senior-engineered Quarkus backend on Java 25 LTS and Maven 3 that
implements the adapted Symphony-for-Trello specification in SPEC.md; work as a senior Quarkus
engineer driven by that spec.
Package manager: Maven via ./mvnw (do not invoke a system Maven). The few Node-based tools
use pnpm at the exact packageManager version, installed as CONTRIBUTING.md Development Setup
says; do not add a package.json only to run a JavaScript CLI.
Before creating any commit that may be pushed, verify the configured author against the
authenticated GitHub account even when the checkout already carries an author-verification marker.
Before every push, inspect the author of every outgoing commit. Never publish generic or test
identities such as Test User or example.invalid; follow the commit and push skills to correct a
feature branch, and stop on the default branch unless the maintainer explicitly authorizes a history
rewrite.
Every commit that reaches main is one cohesive change, on every pull request, outside
contributors' included. Squash a one-change pull request; otherwise clean up its history to one
commit per change and merge it with a rebase merge. Fold each review fix into the commit it fixes,
keeping that commit's author, unless the fix is a cohesive change of its own, and re-check the
Commit History in Main checkboxes after every push. Pull-request review checks the history; see
Committing and pull requests and
Outside contributor pull requests.
Before starting any task, open docs/agents/default-workflow.md and follow it.
When a user corrects a repeatable mistake or states a generally useful working preference, make the durable agent-doc update proactively in the same change; do not wait for the user to explicitly ask for persistence. If the right persistence scope is unclear, ask before finishing. Unless the user explicitly requests a one-off fix, every fix MUST address the general root cause rather than only the immediate symptom. When the original failure can be reproduced safely and within scope, verify it before and after the change to prove that the root-cause fix resolves it. Add a regression test or another durable guardrail proportionate to the failure. If direct reproduction, post-fix verification, or a guardrail is infeasible, document the specific constraint and remaining risk in the handoff. When an installed skill eval issue captures bad or missing behavior, add or update the corresponding temporary repo-local override described in Default workflow and Java style & design preferences. At the point a task makes or explains a deliberate design tradeoff, before implementing or publishing the chosen path, make the decision explicit in the code where possible and check whether Specification & ADR discipline requires an ADR. An implementation, pull-request description, checklist, review reply, or chat summary is not a substitute for a required ADR. Add or update it before the user asks. For repeated literals, scenario-table values, parallel edits that look coupled, and Java collection mutability defaults, apply the centralization and refactoring guidance in Java style & design preferences. When adding or editing tests, apply the AssertJ and test-duplication guidance in Testing, including avoiding assertion loops or assertion streams when AssertJ can express the expectation directly.
Follow the compatibility and migration rules in Compatibility discipline.
Never let a GitHub Actions workflow run or restore anything a pull request controls in a privileged context; follow GitHub Actions security for every workflow change.
For normal code changes, run ./mvnw -q spotless:check verify before finishing (use
spotless:apply first when formatting changed). Static-analysis and lint gates must be clean; give
every suppression the narrowest possible scope and a documented reason.
When asked to "rebase main to the Symphony repo" or similar, follow the upstream Symphony rebase process in Default workflow.
Maintenance
AGENTS.md is intentionally minimal and should not contain topic-specific guidance. See Maintaining agent docs. If you are asked to add detailed guidance here, put it in docs/agents and add a link instead.
More detailed instructions (progressive disclosure)
- Default workflow
- Specification & ADR discipline
- Testing
- Deployment & live verification
- Java style & design preferences
- Static analysis policy
- GitHub Actions workflows: step code lives in
scripts/, not in the workflow file - GitHub Actions security
- Dependency updates
- Documentation & README
- Creating, editing, or triaging GitHub issues: GitHub issue triage;
label triage follows its Triage labels policy and ends with
node scripts/audit-issue-labels.ts, and sweeps answer people's comments and paid-help or bounty offers as that page says - Autonomy & escalation including the Codex review/fix loop
- Private-context redaction
- Maintaining agent docs
- Progressive disclosure source
- Reference provenance
Agent Rules <!-- tessl-managed -->
@.tessl/RULES.md follow the instructions
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.

