synthesis
Skill10 starsChanged 53 days agoMITView on GitHub
Establish the synthesized implementation, its constraints and measurements, and assess them against the task's requirements.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- Celorus/celorus-plugintheir own version, first seen Nov 6, 2026
- marcusrbrown/metacogstheir own version, first seen Nov 1, 2026
- Seth090502/osanwe-publictheir own version, first seen Oct 20, 2026
- yologdev/yoyo-evolvetheir own version, first seen Oct 8, 2026
- richard-kim-79/archora-skillstheir own version, first seen May 15, 2026
- sruthir28/enterprise-ai-skillstheir own version, first seen Apr 14, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- yologdev/yoyo-evolve★ 1.9ksame name
- sruthir28/enterprise-ai-skills★ 136same name
- richard-kim-79/archora-skills★ 49same name
- marcusrbrown/metacogs★ 1same name
- Celorus/celorus-plugin★ 0same name
- Seth090502/osanwe-public★ 0same name
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 repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- brainstormSkill
- design-flowSkill
- env-precheckSkill
- lint-cdcSkill
- power-analysisSkill
- rtl-designSkill
- simulation-planSkill
- simulationSkill
- simulation-triageSkill
- specificationSkill
- timing-analysisSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Write under {workdir}; upstream inputs remain read-only.
- Distinctive: line 33
Check names and divider ratios against RTL.
- Very specific: line 51
Use current requirements with the applicable reports.
“Common” is measured: kept by at least 6 in 10 of the 4 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
Synthesis
Establish the implementation and evidence needed for the current task. Read dispatch.json,
the original intent, requirements, RTL and constraints, and the materials carried into {workdir}. A new dispatch
requires a current conclusion; it does not prescribe a new synthesis. Use measurements whose
implementation, conditions and scope still apply, and perform the work the remaining questions need.
Write under {workdir}; upstream inputs remain read-only. The framework carries previously
published setup and products into this directory, without the old verdict. Keep useful analysis
and comparison material in evidence/.
Tools
<skill> is this skill's directory. Setup, execution and judgment are independent operations:
python3 <skill>/scripts/synthesis/__main__.py bootstrap --workdir {workdir}Bootstrap refreshes scripts/rtl_load.tcl and constraints.sdc from upstream inputs. It installs
missing config.tcl, constraints.local.sdc and driver scripts, preserving existing copies of these
editable files. manifest.module supplies the top; rtl-files.json supplies source order and include
paths. Adapt scripts/dc_run.tcl for the required calculation. The supplied script maps
RTL using DC-Ultra; examining an existing mapped implementation may require different tool work.
Render the design's declared generated clocks and timing exceptions from
constraint-annotations.json into constraints.local.sdc, which follows constraints.sdc.
Check names and divider ratios against RTL. Record the engineering basis for exceptions, local
clock uncertainty and library-specific IO settings alongside the SDC commands.
Set the actual LIB_DB and WIRE_LOAD_MODEL (a library model or none) in config.tcl, then run:
cd {workdir}
python3 <skill>/scripts/synthesis/__main__.py run --workdir .run computes in a temporary directory using the current setup. Scripts write outputs relative
to out/ and reports/ in their working directory, including supplementary reports, and read
inputs at explicit paths. Existing reports are withdrawn;
out/ and reports/ are replaced only after successful, readable output. The log and failed
calculation remain available until retry. Wait for completion and inspect warnings and conditions.
Judge and close
Use current requirements with the applicable reports. finalize compares area_um2 and
timing_slack_ns targets; setup slack comes from the high-precision timing_setup.rpt, whose
report_units output supplies the conversion to ns. QOR checks consistency. Refresh a report lacking
units or sufficient precision from the applicable
implementation and constraints; that alone does not require resynthesis. Declare evidence-based
judgments for synthesis rows without numerical targets, including library-based equivalent-gate units.
Numeric targets cannot be overridden by declarations.
python3 <skill>/scripts/synthesis/__main__.py finalize --workdir {workdir} \
[--requirements '[{"id":"R-1","met":true,"actual":"measured value","measured":"report and scope"}]'] \
[--fail-reason "unresolved cause"] [--fix-owner <rule>]out/ delivers the netlist, matching nonempty SDC/SDF and any supporting files they require.
Consumers track this complete directory. Keep reports and explanatory evidence outside out/;
changes to the package require reassessment, not necessarily new computation. Use --fail-reason for invalid or incomplete work even when numbers
exist. Name the repair owner from the defect, including this stage for its own work. Acceptance
changes follow the user's actual authorization.
Finalize withdraws the previous result before judging. Exit 0 means a pass/fail result was written;
return STATUS: DONE. A nonzero exit means no result was produced; resolve the cause or return
STATUS: BLOCKED <cause>. A failing requirement is a conclusion, not a tool execution failure.
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.

